We propose a novel self-supervised Video Object Segmentation (VOS) approach that strives to achieve better visual correspondence across frames and object-background discriminability for accurate object segmentation. Distinct from previous self-supervised VOS methods, our approach is based on a learning loss formulation that takes into account both object and background information to ensure object-background discriminability, rather than using only object appearance. The objective function comprises cutout-based reconstruction (cutout region represents part of a frame, whose pixels are replaced with some constant values) and tag prediction loss terms. The cutout-based reconstruction term utilizes a simple cutout scheme to learn the pixel-wise correspondence between the current and previous frames in order to reconstruct the original current frame with added cutout region in it. The introduced cutout patch guides the model to focus on the reappearance of scene parts , thereby implicitly equipping the model to address occlusion-based scenarios Next, the tag prediction term encourages object-background separability by grouping tags of all pixels in the cutout region that are similar, while separating them from the tags of the rest of the reconstructed pixels. Additionally, we introduce a zoom-in scheme that addresses the problem of small object segmentation by capturing fine structural information at multiple scales. Our proposed approach, termed CT-VOS, achieves state-of-the-art results on two challenging benchmarks: DAVIS-2017 and Youtube-VOS. A detailed ablation showcases the importance of the proposed loss formulation to effectively establish correspondences, object-background discriminability, and the impact of our zoom-in scheme to accurately segment small-sized objects.

The synthesis of high-resolution remote sensing images based on text descriptions has great potential in many practical application scenarios. Although deep neural networks have achieved great success in many important remote sensing tasks, generating realistic remote sensing images from text descriptions is still very difficult. To address this challenge, we propose a novel text-to-image modern Hopfield network (Txt2Img-MHN). The main idea of Txt2Img-MHN is to conduct hierarchical prototype learning on both text and image embeddings with modern Hopfield layers. Instead of directly learning concrete but highly diverse text-image joint feature representations for different semantics, Txt2Img-MHN aims to learn the most representative prototypes from text-image embeddings, achieving a coarse-to-fine learning strategy. These learned prototypes can then be utilized to represent more complex semantics in the text-to-image generation task. To better evaluate the realism and semantic consistency of the generated images, we further conduct zero-shot classification on real remote sensing data using the classification model trained on synthesized images. Despite its simplicity, we find that the overall accuracy in the zero-shot classification may serve as a good metric to evaluate the ability to generate an image from text. Extensive experiments on the benchmark remote sensing text-image dataset demonstrate that the proposed Txt2Img-MHN can generate more realistic remote sensing images than existing methods. Code and pre-trained models are available online ( https://github.com/YonghaoXu/Txt2Img-MHN ).

Following unprecedented success on the natural language tasks, Transformers have been successfully applied to several computer vision problems, achieving state-of-the-art results and prompting researchers to reconsider the supremacy of convolutional neural networks (CNNs) as de facto operators. Capitalizing on these advances in computer vision, the medical imaging field has also witnessed growing interest for Transformers that can capture global context compared to CNNs with local receptive fields. Inspired from this transition, in this survey, we attempt to provide a comprehensive review of the applications of Transformers in medical imaging covering various aspects, ranging from recently proposed architectural designs to unsolved issues. Specifically, we survey the use of Transformers in medical image segmentation, detection, classification, restoration, synthesis, registration, clinical report generation, and other tasks. In particular, for each of these applications, we develop taxonomy, identify application-specific challenges as well as provide insights to solve them, and highlight recent trends. Further, we provide a critical discussion of the fields current state as a whole, including the identification of key challenges, open problems, and outlining promising future directions. We hope this survey will ignite further interest in the community and provide researchers with an up-to-date reference regarding applications of Transformer models in medical imaging. Finally, to cope with the rapid development in this field, we intend to regularly update the relevant latest papers and their open-source implementations at https://github.com/fahadshamshad/awesome-transformers-in-medical-imaging.

Deep Convolution Neural Networks (CNNs) can easily be fooled by subtle, imperceptible changes to the input images. To address this vulnerability, adversarial training creates perturbation patterns and includes them in the training set to robustify the model. In contrast to existing adversarial training methods that only use class-boundary information (e.g., using a cross-entropy loss), we propose to exploit additional information from the feature space to craft stronger adversaries that are in turn used to learn a robust model. Specifically, we use the style and content information of the target sample from another class, alongside its class-boundary information to create adversarial perturbations. We apply our proposed multi-task objective in a deeply supervised manner, extracting multi-scale feature knowledge to create maximally separating adversaries. Subsequently, we propose a max-margin adversarial training approach that minimizes the distance between source image and its adversary and maximizes the distance between the adversary and the target image. Our adversarial training approach demonstrates strong robustness compared to state-of-the-art defenses, generalizes well to naturally occurring corruptions and data distributional shifts, and retains the models accuracy on clean examples.

We propose a novel approach to translate unpaired contrast computed tomography (CT) scans to noncontrast CT scans and the other way around. Solving this task has two important applications: (i) to automatically generate contrast CT scans for patients for whom injecting contrast substance is not an option, and (ii) to enhance the alignment between contrast and non-contrast CT by reducing the differences induced by the contrast substance before registration.Our approach is based on cycle-consistent generative adversarial convolutional transformers, for short, CyTran. Our neural model can be trained on unpaired images, due to the integration of a multi-level cycleconsistency loss. Aside from the standard cycle-consistency loss applied at the image level, we propose to apply additional cycle-consistency losses between intermediate feature representations, which enforces the model to be cycle-consistent at multiple representations levels, leading to superior results. To deal with high-resolution images, we design a hybrid architecture based on convolutional and multi-head attention layers. In addition, we introduce a novel data set, Coltea-Lung-CT-100W, containing 100 3D triphasic lung CT scans (with a total of 37,290 images) collected from 100 female patients (there is one examination per patient). Each scan contains three phases (non-contrast, early portal venous, and late arterial), allowing us to perform experiments to compare our novel approach with state-of-the-art methods for image style transfer.Our empirical results show that CyTran outperforms all competing methods. Moreover, we show that CyTran can be employed as a preliminary step to improve a state-of-the-art medical image alignment method. We release our novel model and data set as open source at: https://github.com/ristea/cycletransformer.Our qualitative and subjective human evaluations reveal that CyTran is the only approach that does not introduce visual artifacts during the translation process. We believe this is a key advantage in our application domain, where medical images need to precisely represent the scanned body parts. (c) 2023 Elsevier B.V. All rights reserved.

Video instance segmentation is one of the core problems in computer vision. Formulating a purely learning-based method, which models the generic track management required to solve the video instance segmentation task, is a highly challenging problem. In this work, we propose a novel learning framework where the entire video instance segmentation problem is modeled jointly. To this end, we design a graph neural network that in each frame jointly processes all detections and a memory of previously seen tracks. Past information is considered and processed via a recurrent connection. We demonstrate the effectiveness of the proposed approach in comprehensive experiments. Our approach operates online at over 25 FPS and obtains 16.3 AP on the challenging OVIS benchmark, setting a new state-of-the-art. We further conduct detailed ablative experiments that validate the different aspects of our approach. Code is available at https://github.com/emibr948/RGNNVIS-PlusPlus.

Astounding results from Transformer models on natural language tasks have intrigued the vision community to study their application to computer vision problems. Among their salient benefits, Transformers enable modeling long dependencies between input sequence elements and support parallel processing of sequence as compared to recurrent networks, e.g., Long short-term memory. Different from convolutional networks, Transformers require minimal inductive biases for their design and are naturally suited as set-functions. Furthermore, the straightforward design of Transformers allows processing multiple modalities (e.g., images, videos, text, and speech) using similar processing blocks and demonstrates excellent scalability to very large capacity networks and huge datasets. These strengths have led to exciting progress on a number of vision tasks using Transformer networks. This survey aims to provide a comprehensive overview of the Transformer models in the computer vision discipline. We start with an introduction to fundamental concepts behind the success of Transformers, i.e., self-attention, large-scale pre-training, and bidirectional feature encoding. We then cover extensive applications of transformers in vision including popular recognition tasks (e.g., image classification, object detection, action recognition, and segmentation), generative modeling, multi-modal tasks (e.g., visual-question answering, visual reasoning, and visual grounding), video processing (e.g., activity recognition, video forecasting), low-level vision (e.g., image super-resolution, image enhancement, and colorization), and three-dimensional analysis (e.g., point cloud classification and segmentation). We compare the respective advantages and limitations of popular techniques both in terms of architectural design and their experimental value. Finally, we provide an analysis on open research directions and possible future works. We hope this effort will ignite further interest in the community to solve current challenges toward the application of transformer models in computer vision.

In a real-world setting, object instances from new classes can be continuously encountered by object detectors. When existing object detectors are applied to such scenarios, their performance on old classes deteriorates significantly. A few efforts have been reported to address this limitation, all of which apply variants of knowledge distillation to avoid catastrophic forgetting. We note that although distillation helps to retain previous learning, it obstructs fast adaptability to new tasks, which is a critical requirement for incremental learning. In this pursuit, we propose a meta-learning approach that learns to reshape model gradients, such that information across incremental tasks is optimally shared. This ensures a seamless information transfer via a meta-learned gradient preconditioning that minimizes forgetting and maximizes knowledge transfer. In comparison to existing meta-learning methods, our approach is task-agnostic, allows incremental addition of new-classes and scales to high-capacity models for object detection. We evaluate our approach on a variety of incremental learning settings defined on PASCAL-VOC and MS COCO datasets, where our approach performs favourably well against state-of-the-art methods. Code and trained models: https://github.com/JosephKJ/iOD.

We propose a fast single-stage method for both image and video instance segmentation, called SipMask, that preserves the instance spatial information by performing multiple sub-region mask predictions. The main module in our method is a light-weight spatial preservation (SP) module that generates a separate set of spatial coefficients for the sub-regions within a bounding-box, enabling a better delineation of spatially adjacent instances. To better correlate mask prediction with object detection, we further propose a mask alignment weighting loss and a feature alignment scheme. In addition, we identify two issues that impede the performance of single-stage instance segmentation and introduce two modules, including a sample selection scheme and an instance refinement module, to address these two issues. Experiments are performed on both image instance segmentation dataset MS COCO and video instance segmentation dataset YouTube-VIS. On MS COCO test-dev set, our method achieves a state-of-the-art performance. In terms of real-time capabilities, it outperforms YOLACT by a gain of 3.0% (mask AP) under the similar settings, while operating at a comparable speed. On YouTube-VIS validation set, our method also achieves promising results. The source code is available at https://github.com/JialeCao001/SipMask.

Dual-energy computed tomography (CT) can be used in radiotherapy treatment planning for the calculation of absorbed dose distributions. The aim of this work is to evaluate whether there is room for improvement in the accuracy of the Monoenergetic Plus algorithm by Siemens Healthineers. A Siemens SOMATOM Force scanner was used to scan a cylindrical polymethyl methacrylate phantom with four rod-inserts made of different materials. Images were reconstructed using ADMIRE and processed with Monoenergetic Plus. The resulting CT numbers were compared with tabulated values and values simulated by the proof-of-a-concept algorithm DIRA developed by the authors. Both the Monoenergetic Plus and DIRA algorithms performed well; the accuracy of attenuation coefficients was better than about ±1% at the energy of 70 keV. Compared with DIRA, the worse performance of Monoenergetic Plus was caused by its (i) two-material decomposition to iodine and water and (ii) imperfect suppression of the beam hardening artifact in ADMIRE.

The choice of the material base to which the material decomposition is performed in dual-energy computed tomography may affect the quality of reconstructed images. The aim of this work is to investigate how the commonly used bases (water, bone), (water, iodine) and (photoelectric effect, Compton scattering) affect the reconstructed linear attenuation coefficient in the case of the Alvarez–Macovski method. The performance of this method is also compared with the performance of the Dual-energy Iterative Reconstruction Algorithm (DIRA). In both cases, the study is performed using simulations. The results show that the Alvarez–Macovski method produced artefacts when iodine was present in the phantom together with human tissues since this method can only work with one doublet. It was shown that these artefacts could be avoided with DIRA using the (water, bone) doublet for tissues and the (water, iodine) doublet for the iodine solution.

Nowadays, airline ticket prices can vary dynamically and significantly for the same flight, even for nearby seats within the same cabin. Customers are seeking to get the lowest price while airlines are trying to keep their overall revenue as high as possible and maximize their profit. Airlines use various kinds of computational techniques to increase their revenue such as demand prediction and price discrimination. From the customer side, two kinds of models are proposed by different researchers to save money for customers: models that predict the optimal time to buy a ticket and models that predict the minimum ticket price. In this paper, we present a review of customer side and airlines side prediction models. Our review analysis shows that models on both sides rely on limited set of features such as historical ticket price data, ticket purchase date and departure date. Features extracted from external factors such as social media data and search engine query are not considered. Therefore, we introduce and discuss the concept of using social media data for ticket/demand prediction. (c) 2019 The Authors. Production and hosting by Elsevier B.V. on behalf of King Saud University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Automatic recognition of an online series of unsegmented actions requires a method for segmentation that determines when an action starts and when it ends. In this paper, a novel approach for recognizing unsegmented actions in online test experiments is proposed. The method uses self-organizing neural networks to build a three-layer cognitive architecture. The unique features of an action sequence are represented as a series of elicited key activations by the first-layer self-organizing map. An average length of a key activation vector is calculated for all action sequences in a training set and adjusted in learning trials to generate input patterns to the second-layer self-organizing map. The pattern vectors are clustered in the second layer, and the clusters are then labeled by an action identity in the third layer neural network. The experiment results show that although the performance drops slightly in online experiments compared to the offline tests, the ability of the proposed architecture to deal with the unsegmented action sequences as well as the online performance makes the system more plausible and practical in real-case scenarios.

In many computer vision applications, one acquires images of planar surfaces from two different vantage points. One can use a projective transformation to map pixel coordinates associated with a particular planar surface from one image to another. The transformation, called a homography, can be represented by a unique, to within a scale factor, 3 × 3 matrix. One requires a different homography matrix, scale differences apart, for each planar surface whose two images one wants to relate. However, a collection of homography matrices forms a valid set only if the matrices satisfy consistency constraints implied by the rigidity of the motion and the scene. We explore what it means for a set of homography matrices to be compatible and show that two seemingly disparate definitions are in fact equivalent. Our insight lays the theoretical foundations upon which the derivation of various sets of homography consistency constraints can proceed.

Facial expression recognition applications demand accurate and fast algorithms that can run in real time on platforms with limited computational resources. We propose an algorithm that bridges the gap between precise but slow methods and fast but less precise methods. The algorithm combines gentle boost decision trees and neural networks. The gentle boost decision trees are trained to extract highly discriminative feature vectors (local binary features) for each basic facial expression around distinct facial landmark points. These sparse binary features are concatenated and used to jointly optimize facial expression recognition through a shallow neural network architecture. The joint optimization improves the recognition rates of difficult expressions such as fear and sadness. Furthermore, extensive experiments in both within- and cross-database scenarios have been conducted on relevant benchmark data sets for facial expression recognition: CK+, MMI, JAFFE, and SFEW 2.0. The proposed method (LBF-NN) compares favorably with state-of-the-art algorithms while achieving an order of magnitude improvement in execution time.

This paper investigates the problem of position estimation of unmanned surface vessels (USVs) operating in coastal areas or in the archipelago. We propose a position estimation method where the horizon line is extracted in a 360 degrees panoramic image around the USV. We design a convolutional neural network (CNN) architecture to determine an approximate horizon line in the image and implicitly determine the camera orientation (the pitch and roll angles). The panoramic image is warped to compensate for the camera orientation and to generate an image from an approximately level camera. A second CNN architecture is designed to extract the pixelwise horizon line in the warped image. The extracted horizon line is correlated with digital elevation model data in the Fourier domain using a minimum output sum of squared error correlation filter. Finally, we determine the location of the maximum correlation score over the search area to estimate the position of the USV. Comprehensive experiments are performed in field trials conducted over 3 days in the archipelago. Our approach provides excellent results by achieving robust position estimates with global positioning system (GPS)-level accuracy in previously unvisited test areas.

Mobile manipulation robots have great potential for roles in support of rescuers on disaster-response missions. Robots can operate in places too dangerous for humans and therefore can assist in accomplishing hazardous tasks while their human operators work at a safe distance. We developed a disaster-response system that consists of the highly flexible Centauro robot and suitable control interfaces, including an immersive telepresence suit and support-operator controls offering different levels of autonomy.

This paper revisits the problem of continuous-time structure from motion, and introduces a number of extensions that improve convergence and efficiency. The formulation with a C2-continuous spline for the trajectory naturally incorporates inertial measurements, as derivatives of the sought trajectory. We analyze the behavior of split spline interpolation on SO(3) and on R3, and a joint spline on SE(3), and show that the latter implicitly couples the direction of translation and rotation. Such an assumption can make good sense for a camera mounted on a robot arm, but not for hand-held or body-mounted cameras. Our experiments in the Spline Fusion framework show that a split spline on R3andSO(3) is preferable over an SE(3) spline in all tested cases. Finally, we investigate the problem of landmark reprojection on rolling shutter cameras, and show that the tested reprojection methods give similar quality, whereas their computational load varies by a factor of two.

Current best local descriptors are learned on a large data set of matching and non-matching keypoint pairs. However, data of this kind are not always available, since the detailed keypoint correspondences can be hard to establish. On the other hand, we can often obtain labels for pairs of keypoint bags. For example, keypoint bags extracted from two images of the same object under different views form a matching pair, and keypoint bags extracted from images of different objects form a non-matching pair. On average, matching pairs should contain more corresponding keypoints than non-matching pairs. We describe an end-to-end differentiable architecture that enables the learning of local keypoint descriptors from such weakly labeled data. In addition, we discuss how to improve the method by incorporating the procedure of mining hard negatives. We also show how our approach can be used to learn convolutional features from unlabeled video signals and 3D models.

We combine the near-sensor image processing concept with address-event representation leading to an intensity-ranking image sensor (IRIS) and show the benefits of using this type of sensor for image classification. The functionality of IRIS is to output pixel coordinates (X and Y values) continuously as each pixel has collected a certain number of photons. Thus, the pixel outputs will be automatically intensity ranked. By keeping track of the timing of these events, it is possible to record the full dynamic range of the image. However, in many cases, this is not necessary-the intensity ranking in itself gives the needed information for the task at hand. This paper describes techniques for classification and proposes a particular variant (groves) that fits the IRIS architecture well as it can work on the intensity rankings only. Simulation results using the CIFAR-10 dataset compare the results of the proposed method with the more conventional ferns technique. It is concluded that the simultaneous sensing and classification obtainable with the IRIS sensor yields both fast (shorter than full exposure time) and processing-efficient classification.

Color description is one of the fundamental problems of image understanding. One of the popular ways to represent colors is by means of color names. Most existing work on color names focuses on only the eleven basic color terms of the English language. This could be limiting the discriminative power of these representations, and representations based on more color names are expected to perform better. However, there exists no clear strategy to choose additional color names. We collect a dataset of 28 additional color names. To ensure that the resulting color representation has high discriminative power we propose a method to order the additional color names according to their complementary nature with the basic color names. This allows us to compute color name representations with high discriminative power of arbitrary length. In the experiments we show that these new color name descriptors outperform the existing color name descriptor on the task of visual tracking, person re-identification and image classification.

Most methods that address computer vision prob-lems require powerful visual features. Many successfulapproaches apply techniques motivated from nonparametricstatistics. The channel representation provides a frameworkfornonparametricdistributionrepresentation.Althoughearlywork has focused on a signal processing view of the rep-resentation, the channel representation can be interpretedin probabilistic terms, e.g., representing the distribution oflocal image orientation. In this paper, a variety of approxi-mative channel-based algorithms for probabilistic problemsare presented: a novel efficient algorithm for density recon-struction, a novel and efficient scheme for nonlinear griddingof densities, and finally a novel method for estimating Copuladensities. The experimental results provide evidence that byrelaxing the requirements for exact solutions, efficient algo-rithms are obtained

Purpose: To develop and evaluate-in a proof-of-concept configuration-a novel iterative reconstruction algorithm (DIRA) for quantitative determination of elemental composition of patient tissues for application to brachytherapy with low energy (amp;lt; 50 keV) photons and proton therapy. Methods: DIRA was designed as a model-based iterative reconstruction algorithm, which uses filtered backprojection, automatic segmentation and multimaterial tissue decomposition. The evaluation was done for a phantom derived from the voxelized ICRP 110 male phantom. Soft tissues were decomposed to the lipid, protein and water triplet, bones were decomposed to the compact bone and bone marrow doublet. Projections were derived using the Drasim simulation code for an axial scanning configuration resembling a typical DECT (dual-energy CT) scanner with 80 kV and Sn140 kV x-ray spectra. The iterative loop produced mono-energetic images at 50 and 88 keV without beam hardening artifacts. Different noise levels were considered: no noise, a typical noise level in diagnostic imaging and reduced noise level corresponding to tenfold higher doses. An uncertainty analysis of the results was performed using type A and B evaluations. The two approaches were compared. Results: Linear attenuation coefficients averaged over a region were obtained with relative errors less than 0.5% for all evaluated regions. Errors in average mass fractions of the three-material decomposition were less than 0.04 for no noise and reduced noise levels and less than 0.11 for the typical noise level. Mass fractions of individual pixels were strongly affected by noise, which slightly increased after the first iteration but subsequently stabilized. Estimates of uncertainties in mass fractions provided by the type B evaluation differed from the type A estimates by less than 1.5% for most cases. The algorithm was fast, the results converged after 5 iterations. The algorithmic complexity of forward polyenergetic projection calculation was much reduced by using material doublets and triplets. Conclusions: The simulations indicated that DIRA is capable of determining elemental composition of tissues, which are needed in brachytherapy with low energy (amp;lt; 50 keV) photons and proton therapy. The algorithm provided quantitative monoenergetic images with beam hardening artifacts removed. Its convergence was fast, image sharpness expressed via the modulation transfer function was maintained, and image noise did not increase with the number of iterations. c 2017 American Association of Physicists in Medicine

Mammalian herbivores have important top-down effects on ecological processes and landscapes by generating vegetation changes through grazing and trampling. For free-ranging herbivores on large landscapes, trampling is an important ecological factor. However, whereas grazing is widely studied, low-intensity trampling is rarely studied and quantified. The cold-adapted northern tundra reindeer (Rangifer tarandus) is a wide-ranging keystone herbivore in large open alpine and Arctic ecosystems. Reindeer may largely subsist on different species of slow-growing ground lichens, particularly in winter. Lichen grows in dry, snow-poor habitats with frost. Their varying elasticity makes them suitable for studying trampling. In replicated factorial experiments, high-resolution 3D laser scanning was used to quantify lichen volume loss from trampling by a reindeer hoof. Losses were substantial, that is, about 0.3 dm3 per imprint in dry thick lichen, but depended on type of lichen mat and humidity. Immediate trampling volume loss was about twice as high in dry, compared to humid thin (2–3 cm), lichen mats and about three times as high in dry vs. humid thick (6–8 cm) lichen mats, There was no significant difference in volume loss between 100% and 50% wetted lichen. Regained volume with time was insignificant for dry lichen, whereas 50% humid lichen regained substantial volumes, and 100% humid lichen regained almost all lost volume, and mostly within 10–20 min. Reindeer trampling may have from near none to devastating effects on exposed lichen forage. During a normal week of foraging, daily moving 5 km across dry 6- to 8-cm-thick continuous lichen mats, one adult reindeer may trample a lichen volume corresponding to about a year's supply of lichen. However, the lichen humidity appears to be an important factor for trampling loss, in addition to the extent of reindeer movement.

In this work, we introduce a novel tensor-based functional for targeted image enhancement and denoising. Via explicit regularization, our formulation incorporates application-dependent and contextual information using first principles. Few works in literature treat variational models that describe both application-dependent information and contextual knowledge of the denoising problem. We prove the existence of a minimizer and present results on tensor symmetry constraints, convexity, and geometric interpretation of the proposed functional. We show that our framework excels in applications where nonlinear functions are present such as in gamma correction and targeted value range filtering. We also study general denoising performance where we show comparable results to dedicated PDE-based state-of-the-art methods.

1 Migratory songbirds carry an inherited capacity to migrate several thousand kilometers each year crossing continental landmasses and barriers between distant breeding sites and wintering areas. How individual songbirds manage with extreme precision to find their way is still largely unknown. The functional characteristics of biological compasses used by songbird migrants has mainly been investigated by recording the birds directed migratory activity in circular cages, so-called Emlen funnels. This method is 50 years old and has not received major updates over the past decades. The aim of this work was to compare the results from newly developed digital methods with the established manual methods to evaluate songbird migratory activity and orientation in circular cages. 2 We performed orientation experiments using the European robin (Erithacus rubecula) using modified Emlen funnels equipped with thermal paper and simultaneously recorded the songbird movements from above. We evaluated and compared the results obtained with five different methods. Two methods have been commonly used in songbirds orientation experiments; the other three methods were developed for this study and were based either on evaluation of the thermal paper using automated image analysis, or on the analysis of videos recorded during the experiment. 3 The methods used to evaluate scratches produced by the claws of birds on the thermal papers presented some differences compared with the video analyses. These differences were caused mainly by differences in scatter, as any movement of the bird along the sloping walls of the funnel was recorded on the thermal paper, whereas video evaluations allowed us to detect single takeoff attempts by the birds and to consider only this behavior in the orientation analyses. Using computer vision, we were also able to identify and separately evaluate different behaviors that were impossible to record by the thermal paper. 4 The traditional Emlen funnel is still the most used method to investigate compass orientation in songbirds under controlled conditions. However, new numerical image analysis techniques provide a much higher level of detail of songbirds migratory behavior and will provide an increasing number of possibilities to evaluate and quantify specific behaviors as new algorithms will be developed.

Advanced model-based iterative reconstruction algorithms in quantitative computed tomography (CT) perform automatic segmentation of tissues to estimate material properties of the imaged object. Compared with conventional methods, these algorithms may improve quality of reconstructed images and accuracy of radiation treatment planning. Automatic segmentation of tissues is, however, a difficult task. The aim of this work was to develop and evaluate an algorithm that automatically segments tissues in CT images of the male pelvis. The newly developed algorithm (MK2014) combines histogram matching, thresholding, region growing, deformable model and atlas-based registration techniques for the segmentation of bones, adipose tissue, prostate and muscles in CT images. Visual inspection of segmented images showed that the algorithm performed well for the five analysed images. The tissues were identified and outlined with accuracy sufficient for the dual-energy iterative reconstruction algorithm whose aim is to improve the accuracy of radiation treatment planning in brachytherapy of the prostate.

Action recognition in still images is a challenging problem in computer vision. To facilitate comparative evaluation independently of person detection, the standard evaluation protocol for action recognition uses an oracle person detector to obtain perfect bounding box information at both training and test time. The assumption is that, in practice, a general person detector will provide candidate bounding boxes for action recognition. In this paper, we argue that this paradigm is suboptimal and that action class labels should already be considered during the detection stage. Motivated by the observation that body pose is strongly conditioned on action class, we show that: 1) the existing state-of-the-art generic person detectors are not adequate for proposing candidate bounding boxes for action classification; 2) due to limited training examples, the direct training of action-specific person detectors is also inadequate; and 3) using only a small number of labeled action examples, the transfer learning is able to adapt an existing detector to propose higher quality bounding boxes for subsequent action classification. To the best of our knowledge, we are the first to investigate transfer learning for the task of action-specific person detection in still images. We perform extensive experiments on two benchmark data sets: 1) Stanford-40 and 2) PASCAL VOC 2012. For the action detection task (i.e., both person localization and classification of the action performed), our approach outperforms methods based on general person detection by 5.7% mean average precision (MAP) on Stanford-40 and 2.1% MAP on PASCAL VOC 2012. Our approach also significantly outperforms the state of the art with a MAP of 45.4% on Stanford-40 and 31.4% on PASCAL VOC 2012. We also evaluate our action detection approach for the task of action classification (i.e., recognizing actions without localizing them). For this task, our approach, without using any ground-truth person localization at test time, outperforms on both data sets state-of-the-art methods, which do use person locations.

An image mosaic is an assembly of a large number of small images, usually called tiles, taken from a specific dictionary/codebook. When viewed as a whole, the appearance of a single large image emerges, i.e. each tile approximates a small block of pixels. ASCII art is a related (and older) graphic design technique for producing images from printable characters. Although automatic procedures for both of these visualization schemes have been studied in the past, some are computationally heavy and cannot offer real-time and interactive performance. We propose an algorithm able to reproduce the quality of existing non-photorealistic rendering techniques, in particular ASCII art and image mosaics, obtaining large performance speed-ups. The basic idea is to partition the input image into a rectangular grid and use a decision tree to assign a tile from a pre-determined codebook to each cell. Our implementation can process video streams from webcams in real time and it is suitable for modestly equipped devices. We evaluate our technique by generating the renderings of a variety of images and videos, with good results. The source code of our engine is publicly available.

Describing textures is a challenging problem in computer vision and pattern recognition. The classification problem involves assigning a category label to the texture class it belongs to. Several factors such as variations in scale, illumination and viewpoint make the problem of texture description extremely challenging. A variety of histogram based texture representations exists in literature. However, combining multiple texture descriptors and assessing their complementarity is still an open research problem. In this paper, we first show that combining multiple local texture descriptors significantly improves the recognition performance compared to using a single best method alone. This gain in performance is achieved at the cost of high-dimensional final image representation. To counter this problem, we propose to use an information-theoretic compression technique to obtain a compact texture description without any significant loss in accuracy. In addition, we perform a comprehensive evaluation of pure color descriptors, popular in object recognition, for the problem of texture classification. Experiments are performed on four challenging texture datasets namely, KTH-TIPS-2a, KTH-TIPS-2b, FMD and Texture-10. The experiments clearly demonstrate that our proposed compact multi-texture approach outperforms the single best texture method alone. In all cases, discriminative color names outperforms other color features for texture classification. Finally, we show that combining discriminative color names with compact texture representation outperforms state-of-the-art methods by 7.8%,4.3%7.8%,4.3% and 5.0%5.0% on KTH-TIPS-2a, KTH-TIPS-2b and Texture-10 datasets respectively.

Person description is a challenging problem in computer vision. We investigated two major aspects of person description: 1) gender and 2) action recognition in still images. Most state-of-the-art approaches for gender and action recognition rely on the description of a single body part, such as face or full-body. However, relying on a single body part is suboptimal due to significant variations in scale, viewpoint, and pose in real-world images. This paper proposes a semantic pyramid approach for pose normalization. Our approach is fully automatic and based on combining information from full-body, upper-body, and face regions for gender and action recognition in still images. The proposed approach does not require any annotations for upper-body and face of a person. Instead, we rely on pretrained state-of-the-art upper-body and face detectors to automatically extract semantic information of a person. Given multiple bounding boxes from each body part detector, we then propose a simple method to select the best candidate bounding box, which is used for feature extraction. Finally, the extracted features from the full-body, upper-body, and face regions are combined into a single representation for classification. To validate the proposed approach for gender recognition, experiments are performed on three large data sets namely: 1) human attribute; 2) head-shoulder; and 3) proxemics. For action recognition, we perform experiments on four data sets most used for benchmarking action recognition in still images: 1) Sports; 2) Willow; 3) PASCAL VOC 2010; and 4) Stanford-40. Our experiments clearly demonstrate that the proposed approach, despite its simplicity, outperforms state-of-the-art methods for gender and action recognition.

In this article we investigate the problem of human action recognition in static images. By action recognition we intend a class of problems which includes both action classification and action detection (i.e. simultaneous localization and classification). Bag-of-words image representations yield promising results for action classification, and deformable part models perform very well object detection. The representations for action recognition typically use only shape cues and ignore color information. Inspired by the recent success of color in image classification and object detection, we investigate the potential of color for action classification and detection in static images. We perform a comprehensive evaluation of color descriptors and fusion approaches for action recognition. Experiments were conducted on the three datasets most used for benchmarking action recognition in still images: Willow, PASCAL VOC 2010 and Stanford-40. Our experiments demonstrate that incorporating color information considerably improves recognition performance, and that a descriptor based on color names outperforms pure color descriptors. Our experiments demonstrate that late fusion of color and shape information outperforms other approaches on action recognition. Finally, we show that the different color–shape fusion approaches result in complementary information and combining them yields state-of-the-art performance for action classification.

Dosimetric accuracy of radiation treatment planning in brachytherapy depends on knowledge of tissue composition. It has been speculated that soft tissues can be decomposed to water, lipid and protein. The aim of our work is to evaluate the accuracy of such tissue decomposition. Selected abdominal soft tissues, whose average elemental compositions were taken from literature, were decomposed using dual energy computed tomography to water, lipid and protein via the three-material decomposition method. The quality of the decomposition was assessed using relative differences between (i) mass energy absorption and (ii) mass energy attenuation coefficients of the analyzed and approximated tissues. It was found that the relative differences were less than 2% for photon energies larger than 10 keV. The differences were notably smaller than the ones for water as the transport and dose scoring medium. The choice of the water, protein and lipid triplet resulted in negative elemental mass fractions for some analyzed tissues. As negative elemental mass fractions cannot be used in general purpose particle transport computer codes using the Monte Carlo method, other triplets should be used for the decomposition. These triplets may further improve the accuracy of the approximation as the differences were mainly caused by the lack of high-Z materials in the water, protein and lipid triplet.

We propose a novel method for iterative learning of point correspondences between image sequences. Points moving on surfaces in 3D space are projected into two images. Given a point in either view, the considered problem is to determine the corresponding location in the other view. The geometry and distortions of the projections are unknown as is the shape of the surface. Given several pairs of point-sets but no access to the 3D scene, correspondence mappings can be found by excessive global optimization or by the fundamental matrix if a perspective projective model is assumed. However, an iterative solution on sequences of point-set pairs with general imaging geometry is preferable. We derive such a method that optimizes the mapping based on Neyman's chi-square divergence between the densities representing the uncertainties of the estimated and the actual locations. The densities are represented as channel vectors computed with a basis function approach. The mapping between these vectors is updated with each new pair of images such that fast convergence and high accuracy are achieved. The resulting algorithm runs in real-time and is superior to state-of-the-art methods in terms of convergence and accuracy in a number of experiments.

This work proposes an approach to tracking by regression that uses no hard-coded models and no offline learning stage. The Linear Predictor (LP) tracker has been shown to be highly computationally efficient, resulting in fast tracking. Regression tracking techniques tend to require offline learning to learn suitable regression functions. This work removes the need for offline learning and therefore increases the applicability of the technique. The online-LP tracker can simply be seeded with an initial target location, akin to the ubiquitous Lucas-Kanade algorithm that tracks by registering an image template via minimisation. A fundamental issue for all trackers is the representation of the target appearance and how this representation is able to adapt to changes in target appearance over time. The two proposed methods, LP-SMAT and LP-MED, demonstrate the ability to adapt to large appearance variations by incrementally building an appearance model that identifies modes or aspects of the target appearance and associates these aspects to the Linear Predictor trackers to which they are best suited. Experiments comparing and evaluating regression and registration techniques are presented along with performance evaluations favourably comparing the proposed tracker and appearance model learning methods to other state of the art simultaneous modelling and tracking approaches.

In this article we describe a set of canonical transformations of the image spaces that make the description of three-view geometry very simple. The transformations depend on the three-view geometry and the canonically transformed trifocal tensor T' takes the form of a sparse array where 17 elements in well-defined positions are zero, it has a linear relation to the camera matrices and to two of the fundamental matrices, a third order relation to the third fundamental matrix, a second order relation to the other two trifocal tensors, and first order relations to the 10 three-view all-point matching constraints. In this canonical form, it is also simple to determine if the corresponding camera configuration is degenerate or co-linear. An important property of the three canonical transformations of the images spaces is that they are in SO(3). The 9 parameters needed to determine these transformations and the 9 parameters that determine the elements of T' together provide a minimal parameterization of the tensor. It does not have problems with multiple maps or multiple solutions that other parameterizations have, and is therefore simple to use. It also provides an implicit representation of the trifocal internal constraints: the sparse canonical representation of the trifocal tensor can be determined if and only if it is consistent with its internal constraints. In the non-ideal case, the canonical transformation can be determined by solving a minimization problem and a simple algorithm for determining the solution is provided. This allows us to extend the standard linear method for estimation of the trifocal tensor to include a constraint enforcement as a final step, similar to the constraint enforcement of the fundamental matrix.

Experimental evaluation of this extended linear estimation method shows that it significantly reduces the geometric error of the resulting tensor, but on average the algebraic estimation method is even better. For a small percentage of cases, however, the extended linear method gives a smaller geometric error, implying that it can be used as a complement to the algebraic method for these cases.

In this paper, we present a novel scheme for anisotropic diffusion driven by the image autocorrelation function. We show the equivalence of this scheme to a special case of iterated adaptive filtering. By determining the diffusion tensor field from an autocorrelation estimate, we obtain an evolution equation that is computed from a scalar product of diffusion tensor and the image Hessian. We propose further a set of filters to approximate the Hessian on a minimized spatial support. On standard benchmarks, the resulting method performs favorable in many cases, in particular at low noise levels. In a GPU implementation, video real-time performance is easily achieved.

Recent years have seen advances in the estimation of full 6 degree-of-freedom object pose from a single 2D image. These advances have often been presented as a result of, or together with, a new local image feature type. This paper examines how the pose accuracy and recognition robustness for such a system varies with choice of feature type. This is done by evaluating a full 6 degree-of-freedom pose estimation system for 17 different combinations of local descriptors and detectors. The evaluation is done on data sets with photos of challenging 3D objects with simple and complex backgrounds and varying illumination conditions. We examine the performance of the system under varying levels of object occlusion and we find that many features allow considerable object occlusion. From the experiments we can conclude that duplet features, that use pairs of interest points, improve pose estimation accuracy, compared to single point features. Interestingly, we can also show that many features previously used for recognition and wide-baseline stereo are unsuitable for pose estimation, one notable example are the affine covariant features that have been proven quite successful in other applications. The data sets and their ground truths are available on the web to allow future comparison with novel algorithms.

The major goal of the COSPAL project is to develop an artificial cognitive system architecture, with the ability to autonomously extend its capabilities. Exploratory learning is one strategy that allows an extension of competences as provided by the environment of the system. Whereas classical learning methods aim at best for a parametric generalization, i.e., concluding from a number of samples of a problem class to the problem class itself, exploration aims at applying acquired competences to a new problem class, and to apply generalization on a conceptual level, resulting in new models. Incremental or online learning is a crucial requirement to perform exploratory learning. In the COSPAL project, we mainly investigate reinforcement-type learning methods for exploratory learning, and in this paper we focus on the organization of cognitive systems for efficient operation. Learning is used over the entire system. It is organized in the form of four nested loops, where the outermost loop reflects the user-reinforcement-feedback loop, the intermediate two loops switch between different solution modes at symbolic respectively sub-symbolic level, and the innermost loop performs the acquired competences in terms of perception-action cycles. We present a system diagram which explains this process in more detail. We discuss the learning strategy in terms of learning scenarios provided by the user. This interaction between user (teacher) and system is a major difference to classical robotics systems, where the system designer places his world model into the system. We believe that this is the key to extendable robust system behavior and successful interaction of humans and artificial cognitive systems. We furthermore address the issue of bootstrapping the system, and, in particular, the visual recognition module. We give some more in-depth details about our recognition method and how feedback from higher levels is implemented. The described system is however work in progress and no final results are available yet. The available preliminary results that we have achieved so far, clearly point towards a successful proof of the architecture concept.

This paper presents a method that combines shadow detection and a 3D box model including shadow simulation, for estimation of size and position of vehicles. We define a similarity measure between a simulated image of a 3D box, including the box shadow, and a captured image that is classified into background/foreground/shadow. The similarity Measure is used in all optimization procedure to find the optimal box state. It is shown in a number of experiments and examples how the combination shadow detection/simulation improves the estimation compared to just using detection or simulation, especially when the shadow detection or the simulation is inaccurate. We also describe a tracking system that utilizes the estimated 3D boxes, including highlight detection, spatial window instead of a time based window for predicting heading, and refined box size estimates by weighting accumulated estimates depending oil view. Finally, we show example results.

This article introduces a new region based feature for object recognition and image matching. In contrast to many other region based features, this one makes use of colour in the feature extraction stage. We perform experiments on the repeatability rate of the features across scale and inclination angle changes, and show that avoiding to merge regions connected by only a few pixels improves the repeatability. We introduce two voting schemes that allow us to find correspondences automatically, and compare them with respect to the number of valid correspondences they give, and their inlier ratios. We also demonstrate how the matching procedure can be applied to colour correction.

Inspired by the early visual system of many mammalians we consider the construction of-and reconstruction from- an orientation score U_f:R²×S¹→C as a local orientation representation of an image, f:R²→R . The mapping f↦Uf is a wavelet transform W_ψ corresponding to a reducible representation of the Euclidean motion group onto L₂(R²) and oriented wavelet ψ∈L₂(R²) . This wavelet transform is a special case of a recently developed generalization of the standard wavelet theory and has the practical advantage over the usual wavelet approaches in image analysis (constructed by irreducible representations of the similitude group) that it allows a stable reconstruction from one (single scale) orientation score. Since our wavelet transform is a unitary mapping with stable inverse, we directly relate operations on orientation scores to operations on images in a robust manner.

Furthermore, by geometrical examination of the Euclidean motion group G=R²R×T , which is the domain of our orientation scores, we deduce that an operator Φ on orientation scores must be left invariant to ensure that the corresponding operator W−1ψΦW_ψ on images is Euclidean invariant. As an example we consider all linear second order left invariant evolutions on orientation scores corresponding to stochastic processes on G. As an application we detect elongated structures in (medical) images and automatically close the gaps between them.

Finally, we consider robust orientation estimates by means of channel representations, where we combine robust orientation estimation and learning of wavelets resulting in an auto-associative processing of orientation features. Here linear averaging of the channel representation is equivalent to robust orientation estimation and an adaptation of the wavelet to the statistics of the considered image class leads to an auto-associative behavior of the system.

Augmented reality is a growing field, with many diverse applications ranging from TV and film production, to industrial maintenance, medicine, education, entertainment and games. The central idea is to add virtual objects into a real scene, either by displaying them in a see-through head-mounted display, or by superimposing them on an image of the scene captured by a camera. Depending on the application, the added objects might be virtual characters in a TV or film production, instructions for repairing a car engine, or a reconstruction of an archaeological site. For the effect to be believable, the virtual objects must appear rigidly fixed to the real world, which requires the accurate measurement in real-time of the position of the camera or the user-s head. Present technology cannot achieve this without resorting to systems that require a significant infrastructure in the operating environment, severely restricting the range of possible applications.

This paper brings together a novel information representation model for use in signal processing and computer vision problems, with a particular algorithmic development of the Landweber iterative algorithm. The information representation model allows a representation of multiple values for a variable as well as an expression for confidence. Both properties are important for effective computation using multi-level models, where a choice between models will be implementable as part of the optimization process. It is shown that in this way the algorithm can deal with a class of high-dimensional, sparse, and constrained least-squares problems, which arise in various computer vision learning tasks, such as object recognition and object pose estimation. While the algorithm has been applied to the solution of such problems, it has so far been used heuristically. In this paper we describe the properties and some of the peculiarities of the channel representation and optimization, and put them on firm mathematical ground. We consider the optimization a convexly constrained weighted least-squares problem and propose for its solution a projected Landweber method which employs oblique projections onto the closed convex constraint set. We formulate the problem, present the algorithm and work out its convergence properties, including a rate-of-convergence result. The results are put in perspective with currently available projected Landweber methods. An application to supervised learning is described, and the method is evaluated in an experiment involving function approximation, as well as application to transient signals. © 2006 Elsevier Ltd. All rights reserved.

A runtime system for implementation of image processing operations is presented. It is designed for working in a flexible and distributed environment related to the software architecture of a newly developed UAV system. The software architecture can be characterized at a coarse scale as a layered system, with a deliberative layer at the top, a reactive layer in the middle, and a processing layer at the bottom. At a finer scale each of the three levels is decomposed into sets of modules which communicate using CORBA, allowing system development and deployment on the UAV to be made in a highly flexible way. Image processing takes place in a dedicated module located in the process layer, and is the main focus of the paper. This module has been designed as a runtime system for data flow graphs, allowing various processing operations to be created online and on demand by the higher levels of the system. The runtime system is implemented in Java, which allows development and deployment to be made on a wide range of hardware/software configurations. Optimizations for particular hardware platforms have been made using Java's native interface.

 In medical helical cone-beam CT, it is common that the region-of-interest (ROI) is contained inside the helix cylinder, while the complete object is long and extends outside the top and the bottom of the cylinder. This is the Long Object Problem. Analytical reconstruction methods for helical cone-beam CT have been designed to handle this problem. It has been shown that a moderate amount of over-scanning is sufficient for reconstruction of a certain ROI. The over-scanning projection rays travel both through the ROI as well as outside the ROI. This is unfortunate for iterative methods since it seems impossible to compute accurate values for the projection rays which travel partly inside and partly outside the ROI. Therefore, it seems that the useful ROI will diminish for every iteration step. We propose the following solution to the problem. Firstly, we reconstruct volume regions also outside the ROI. These volume regions will certainly be incompletely reconstructed, but our experimental results show that they serve well for projection generation. This is rather counter-intuitive and contradictory to our initial assumptions. Secondly, we use careful extrapolation and masking of projection data. This is not a general necessity, but needed for the chosen iterative algorithm, which includes rebinning and iterative filtered backprojection. Our idea here was to use an approximate reconstruction method which gives cone-beam artifacts and then improve the reconstructed result by iterative filtered backprojection. The experimental results seem very encouraging. The cone-beam artifacts can indeed be removed. Even voxels close to the boundary of the ROI are as well enhanced by the iterative loop as those in the middle of the ROI.

Tensors have become a popular tool for representation of local orientation and can be used also for estimation of velocity. A number of computational approaches have been presented for tensor estimation which, however, are difficult to analyze or compare since there has been no common framework in which analysis or comparisons can be made. In this article, we propose such a framework based on second-order filters and show how it applies to three different methods for tensor estimation. The framework contains a few conditions on the filters which are sufficient to obtain correctly oriented rank one tensors for the case of simple signals. It also allows the derivation of explicit expressions for the variation of the tensor across oriented structures which, e.g., can be used to formulate conditions for phase invariance. (c) 2005 Elsevier B.V. All rights reserved.

In this paper we present a novel method to implement the monogenic scale space on a rectangular domain. The monogenic scale space is a vector valued scale space based on the Poisson scale space, which establishes a sophisticated alternative to the Gaussian scale space. Previous implementations of the monogenic scale space are Fourier transform based, and therefore suffer from the implicit periodicity in case of finite domains. The features of the monogenic scale space, including local amplitude, local phase, local orientation, local frequency, and phase congruency, are much easier to interpret in terms of image features evolving through scale than in the Gaussian case. Furthermore, applying results from harmonic analysis, relations between the features are obtained which improve the understanding of image analysis. As applications, we present a very simple but still accurate approach to image reconstruction from local amplitude and local phase and a method for extracting the evolution of lines and edges through scale.

We introduce a compact coding of image information in terms of local multi-modal image descriptors. This coding allows for an explicit separation of the local image information into different visual sub-modalities: geometric information (orientation) and structural image information (contrast transition and colour). Based on this image representation, we derive a similarity function that compares visual information in each of these sub-modalities. This allows for an investigation of the importance of the different factors for stereo matching on a large data set. From this investigation we conclude that it is the combination of visual modalities that gives the best results. Concrete weights for their relative importance are measured. In addition to these quantitative results, we can demonstrate by our simulations that although our image representation reduces image information by 97% we achieve a matching performance which is comparable to block matching techniques. This shows that our very condensed representation preserves the relevant visual information. © 2004 Elsevier B.V. All rights reserved.

 A method for unrestricted recognition of 3-D objects has been developed. By unrestricted, we imply that the recognition shall be done independently of object position, scale, orientation and pose, against a structured background. It shall not assume any preceding segmentation and allow a reasonable degree of occlusion. The method uses a hierarchy of triplet feature invariants, which are at each level defined by a learning procedure. In the feed-back learning procedure, percepts are mapped upon system states corresponding to manipulation parameters of the object. The method uses a learning architecture employing channel information representation. The paper contains a discussion of how objects can be represented. A structure is proposed to deal with object and contextual properties in a transparent manner.

Estimation of local second-degree variation should be a natural first step in computerized image analysis, just as it seems to be in human vision. A prevailing obstacle is that the second derivatives entangle the three features signal strength (i.e. magnitude or energy), orientation and shape. To disentangle these features we propose a technique where the orientation of an arbitrary pattern f is identified with the rotation required to align the pattern with its prototype p. This is more strictly formulated as solving the derotating equation. The set of all possible prototypes spans the shape-space of second degree variation. This space is one-dimensional for 2Dimages, two-dimensional for 3D-images. The derotation decreases the original dimensionality of the response vector from three to two in the 2D-case and from six to three in the 3D-case, in both cases leaving room only for magnitude and shape in the prototype. The solution to the derotation and a full understanding of the result requires i) mapping the derivatives of the pattern f onto the orthonormal basis of spherical harmonics, and ii) identifying the eigenvalues of the Hessian with the derivatives of the prototype p. But once the shape-space is established the possibilities to put together independent discriminators for magnitude, orientation, and shape are easy and almost limitless.

As welfare diseases become more common all over the world the demand for angiography examinations is increasing rapidly. The development of advanced medical signal processing methods has with few exceptions been concentrated towards CT and MR while traditional contrast based radiology depend on methods developed for ancient photography techniques despite the fact that angiography sequences are generally recorded in digital form. This article presents a new approach for processing of angiography sequences based on advanced image processing methods. The developed algorithm automatically processes angiography sequences containing motion artifacts that cannot be processed by conventional methods like digital subtraction angiography (DSA) and pixel shift due to non uniform motions. The algorithm can in simple terms be described as an ideal pixelshift filter carrying out shifts of different directions and magnitude according to the local motions in the image. In difference to conventional methods it is fully automatic, no mask image needs to be defined and the manual pixelshift operations, which are extremely time consuming, are eliminated. The algorithm is efficient and robust and is designed to run on standard hardware of a powerful workstation which excludes the need for expensive dedicated angiography platforms. Since there is no need to make additional recordings if the patient moves, the patient is exposed to less amount of radiation and contrast fluid. The most exciting benefits by this method are, however, that it opens up new areas for contrast based angiography that are not possible to process with conventional methods e.g. nonuniform motions and multiple layers of moving tissue. Advanced image processing methods provide significantly better image quality and noise suppression but do also provide the means to compute flow velocity and visualize the flow dynamics in the arterial trees by e.g. using color. Initial tests have proven that it is possible to discriminate capillary blood flow from angiography data which opens up interesting possibilities for estimating the blood flow in the heart muscle without use of nuclear methods.

A study of recurrent associative memories with exclusively short-range connections is presented. To increase the capacity, higher order couplings are used. We study capacity and pattern completion ability of networks consisting of units with binary (±1) output. Results show that perfect learning of random patterns is difficult for very short coupling ranges, and that the average expected capacities (allowing small errors) in these cases are much smaller than the theoretical maximum, 2 bits per coupling. However, it is also shown that by choosing ranges longer than certain limit sizes, depending on network size and order, we can come close to the theoretical capacity limit. We indicate that these limit sizes increase very slowly with net size. Thus, couplings to at least 28 and 36 neighbors suffice for second order networks with 400 and 90,000 units, respectively. From simulations it is found that even networks with coupling ranges below the limit size are able to complete input patterns with more than 10% errors. Especially remarkable is the ability to correct inputs with large local errors (part of the pattern is masked). We present a local learning algorithm for heteroassociation in recurrent networks without hidden units. The algorithm is used in a multinet system to improve pattern completion ability on correlated patterns.

The problem of detection of orientation in finite dimensional Euclidean spaces is solved in the least squares sense. In particular, the theory is developed for the case when such orientation computations are necessary at all local neighborhoods of the n-dimensional Euclidean space. Detection of orientation is shown to correspond to fitting an axis or a plane to the Fourier transform of an n-dimensional structure. The solution of this problem is related to the solution of a well-known matrix eigenvalue problem. Moreover, it is shown that the necessary computations can be performed in the spatial domain without actually doing a Fourier transformation. Along with the orientation estimate, a certainty measure, based on the error of the fit, is proposed. Two applications in image analysis are considered: texture segmentation and optical flow. An implementation for 2-D (texture features) as well as 3-D (optical flow) is presented. In the case of 2-D, the method exploits the properties of the complex number field to by-pass the eigenvalue analysis, improving the speed and the numerical stability of the method. The theory is verified by experiments which confirm accurate orientation estimates and reliable certainty measures in the presence of noise. The comparative results indicate that the proposed theory produces algorithms computing robust texture features as well as optical flow. The computations are highly parallelizable and can be used in realtime image analysis since they utilize only elementary functions in a closed form (up to dimension 4) and Cartesian separable convolutions.

Recent physiological research has indicated that the visual system makes use of units responsive to Gabor signals in the analysis of visual stimuli. Such functions effect a tradeoff between pure spatial- and frequency-domain descriptions. The authors explain the use of such representations in vision, considered as a process in inference from the retinal signals to a symbolic description. The appropriate mathematical structure for the inference is that of the subspaces of the signal vector space, a feature which it shares with quantum mechanics. The theory is derived directly from the fundamental constraints on visual inference. It is then shown to be consistent with many of the known properties of the visual system. In particular, a major feature of the inference system-the occurrence of interference effects-has already been observed in visual system operation

A new predictive coder, based on an estimation method which adapts to line and edge features in images, is described. Quantization of the prediction error is performed by a two-level adaptive scheme: an adaptive transform coder, and a threshold coding in both transform and spatial domains. Control information, which determines the behavior of the predictor, is quantized using a simple variable rate technique. The results are improved by pre- and post-filtering using a related noncausal form of the estimator. Acceptable images have been produced in this way at bit rates of less than 0.5 bit/pixel.

In a special radiographic process, ectomography, an image of a slice is produced by simple summation of a set of specially filtered component images, of which each represents one of at least 60 different projections of the object. After being digitized, they are stored, filtered, and summed in a computer. Images representing any slice of any thickness in the object may be produced from the same set of component images. All details within the slice are pictured correctly while details outside are almost completely eliminated.

Radiographic technology has advanced considerably during the last years with the advent of reconstruction techniques allowing visualization of slices through the body. In spite of the advantage of computed tomography compared to conventional radiographic methods, there are still some shortcomings with the method If a different section of the body is desired, another recording has to be made, the width of the dice reconstructed is fixed, and a full 1800 view angle is required.

The advent of new stains for chromosomes has increased the possibility of implementing useful automated chromosome analysis. The case with which chromosomes can now be recognized makes it possible to perform detailed statistical analysis of the chromosomes of an individual. This paper describes methods for assembling chromosome information from several cells in such a way that accidental variations due to preparation, etc. can be eliminated and an undistorted set of characteristics of the chromosome complement can be established. This set of characteristics can then be compared with various references, and statements can be made concerning the relationships between variations in the chromosome complement and genetic traits. These same methods can be employed in multiple-cell karyotyping to circumvent the classical problem of touching and overlapping chromosomes. The methods also allow one to achieve very reliable descriptions of the chromosome complement. The importance of appropriate descriptors of the chromosomes is illustrated.

Under the title "Probabilistic and Biologically Inspired Feature Representations," this text collects a substantial amount of work on the topic of channel representations. Channel representations are a biologically motivated, wavelet-like approach to visual feature descriptors: they are local and compact, they form a computational framework, and the represented information can be reconstructed. The first property is shared with many histogram- and signature-based descriptors, the latter property with the related concept of population codes. In their unique combination of properties, channel representations become a visual Swiss army knife—they can be used for image enhancement, visual object tracking, as 2D and 3D descriptors, and for pose estimation. In the chapters of this text, the framework of channel representations will be introduced and its attributes will be elaborated, as well as further insight into its probabilistic modeling and algorithmic implementation will be given. Channel representations are a useful toolbox to represent visual information for machine learning, as they establish a generic way to compute popular descriptors such as HOG, SIFT, and SHOT. Even in an age of deep learning, they provide a good compromise between hand-designed descriptors and a-priori structureless feature spaces as seen in the layers of deep networks.

Signal Processing for Computer Vision is a unique and thorough treatment of the signal processing aspects of filters and operators for low-level computer vision.

Computer vision has progressed considerably over recent years. From methods only applicable to simple images, it has developed to deal with increasingly complex scenes, volumes and time sequences. A substantial part of this book deals with the problem of designing models that can be used for several purposes within computer vision. These partial models have some general properties of invariance generation and generality in model generation.

Signal Processing for Computer Vision is the first book to give a unified treatment of representation and filtering of higher order data, such as vectors and tensors in multidimensional space. Included is a systematic organisation for the implementation of complex models in a hierarchical modular structure and novel material on adaptive filtering using tensor data representation.

Signal Processing for Computer Vision is intended for final year undergraduate and graduate students as well as engineers and researchers in the field of computer vision and image processing.

The choice of the material base to which the material decomposition is performed in dual-energy computed tomography may affect the quality of reconstructed images. Resulting inaccuracies may lower their diagnostic value, or if the data are used for radiation treatment planning, the accuracy of such plans. The aim of this work is to investigate how the commonly used (water, bone) (WB), (water, iodine) (WI), and (approximate photoelectric effect, Compton scattering) (PC) doublets affect the reconstructed linear attenuation coefficient in the case of the Alvarez–Macovski (AM) method. The performance of this method is also compared to the performance of the dual-energy iterative reconstruction algorithm DIRA. In both cases, the study is performed using simulations.

The results show that the PC and WB doublets accurately predicted the linear attenuation coefficient (LAC) values for human tissues and elements with Z = 1, …, 20, in the 20–150 keV range, though there was a small (<5% discrepancy in the 20–35 keV range. The WI doublet did not represent the tissues as well as PC and WB; the largest discrepancies (>50% in some cases) were in the 20–40 keV range.

LACs reconstructed with the AM and DIRA followed this trend. AM produced artifacts when iodine was present in the phantom together with human tissues since AM can only work with one doublet at a time. It was shown that these artifacts could be avoided with DIRA using different doublets at different spatial positions, i.e., WB for soft and bone tissue and WI for the iodine solution.

Scene reconstruction, i.e. the process of creating a 3D representation (mesh) of some real world scene, has recently become easier with the advent of cheap RGB-D sensors (e.g. the Microsoft Kinect).

Many such sensors use rolling shutter cameras, which produce geometrically distorted images when they are moving. To mitigate these rolling shutter distortions we propose a method that uses an attached gyroscope to rectify the depth scans.We also present a simple scheme to calibrate the relative pose and time synchronization between the gyro and a rolling shutter RGB-D sensor.

For scene reconstruction we use the Kinect Fusion algorithm to produce meshes. We create meshes from both raw and rectified depth scans, and these are then compared to a ground truth mesh. The types of motion we investigate are: pan, tilt and wobble (shaking) motions.

As our method relies on gyroscope readings, the amount of computations required is negligible compared to the cost of running Kinect Fusion.

This chapter is an extension of a paper at the IEEE Workshop on Robot Vision [10]. Compared to that paper, we have improved the rectification to also correct for lens distortion, and use a coarse-to-fine search to find the time shift more quicky.We have extended our experiments to also investigate the effects of lens distortion, and to use more accurate ground truth. The experiments demonstrate that correction of rolling shutter effects yields a larger improvement of the 3D model than correction for lens distortion.

This volume does much more than survey modern advanced color processing. Starting with a historical perspective on ways we have classified color, it sets out the latest numerical techniques for analyzing and processing colors, the leading edge in our search to accurately record and print what we see. The human eye perceives only a fraction of available light wavelengths, yet we live in a multicolor world of myriad shining hues. Colors rich in metaphorical associations make us "purple with rage" or "green with envy" and cause us to "see red." Defining colors has been the work of centuries, culminating in today's complex mathematical coding that nonetheless remains a work in progress: only recently have we possessed the computing capacity to process the algebraic matrices that reproduce color more accurately. With chapters on dihedral color and image spectrometers, this book provides technicians and researchers with the knowledge they need to grasp the intricacies of today's color imaging.

This Chapter is devoted to the cooperation of multiple UAVs for environment perception. First, probabilistic methods for multi-UAV cooperative perception are analyzed. Then, the problem of multi-UAV detection, localization and tracking is described, and local image processing techniques are presented. Then, the Chapter shows two approaches based on the Information Filter and on evidence grid representations.

This volume is a post-event proceedings volume and contains selected papers based on the presentations given, and the lively discussions that ensued, during a seminar held in Dagstuhl Castle, Germany, in October 2003. Co-sponsored by ECVision, the cognitive vision network of excellence, it was organized to further strengthen cooperation between research groups from different countries working in the field of cognitive vision systems.

This paper presents a novel two-frame motion estimation algorithm. The first step is to approximate each neighborhood of both frames by quadratic polynomials, which can be done efficiently using the polynomial expansion transform. From observing how an exact polynomial transforms under translation a method to estimate displacement fields from the polynomial expansion coefficients is derived and after a series of refinements leads to a robust algorithm. Evaluation on the Yosemite sequence shows good results.

Perceptual experiments indicate that corners and curvature are very important features in the process of recognition. This paper presents a new method to efficiently detect rotational symmetries, which describe complex curvature such as corners, circles, star- and spiral patterns. The method is designed to give selective and sparse responses. It works in three steps, first extract local orientation from a gray-scale or color image, second correlate the orientation image with rotational symmetry filters and third let the filter responses inhibit each other in order to get more selective responses, The correlations can be made efficient by separating the 2D-filters into a small number of 1D-filters. These symmetries can serve as feature points at a high abstraction level for use in hierarchical matching structures for 3D-estimation, object recognition, etc.

This chapter is not original, but presents methods for linear classification in the tradition of N. J. Nilsson as well as R. O. Duda and P. E. Hart. Part of the motivation for including this well-known material is to allow the vision structure to be brought to a logical conclusion in which feature properties are combined to form responses or class statements. Another motivation developed here is to display the similarity in structure between convolution operations and linear discriminant functions. This brings all operations for feature extraction and classification to the use of a common component, linear discriminants. This is also illustrated in the form of perceptrons, which allows a transition to the modern theory of neural networks.

This chapter presents a computationally efficient technique for adaptive filtering of n-dimensional  signals.

The approach is based on the local signal description given by the orientation tensor discussed in Chapter 6. The adaptive filter output is synthesized as a tensor-controlled weighted summation of shift-invariant filter outputs. Several examples of adaptive filtering in two and three dimensions are given. The chapter contains original results on the extension of the techniques to n dimensions

This chapter deals with the estimation of local frequency and bandwidth. Local frequency is an important concept which provides an indication of the appropriate range of scales for subsequent analysis. A number of one-dimensional and two-dimensional examples of local frequency and bandwidth estimation are given.

This chapter introduces the use of tensors in estimation of local structure and orientation. The tensor representation is shown to be crucial to unambiguous and continuous representation of local orientation in multiple dimensions. In addition to orientation the tensor representation also conveys the degree and type of local anisotropy. The orientation estimation approach is developed in detail for two, three and four dimensions and is shown to be extendable to higher dimensions. Examples and performance measures are given for processing of images, volumes and time sequences.

This chapter is on Fourier methods, with a particularemphasis on definitions and theorems essential to the understanding offiltering procedures in multi-dimensional spaces. This is a centralissue in computer vision.

This chapter givesan overview of important biological vision mechanisms. Although agreat deal is known about neural processing of visual information,most essential questions about biological vision remain as yetunanswered. Nonetheless, the knowledge available has already provideduseful guidance to the organization of effective machine visionsystems.

We apply the 3D-orientation tensor representation to construct an object tracking algorithm. 2D-line normal flow is estimated by computing the eigenvector associated with the largest eigenvalue of 3D (two spatial dimensions plus time) tensors with a planar structure. Object's true 2D velocity is computed by averaging tensors with consistent normal flows, generating a 3D line represention that corresponds to a 2D point in motion. Flow induced by camera rotation is compensated for by ignoring points with velocity consistent with the ego-rotation. A region-of-interest growing process based on motion consistency generates estimates of object size and position.  Introduction The literature on optical flow estimation is wast. Descriptions and performance studies of a number of different techniques are given in  and the monographs by Fleet  and Jahne. We will only briefly describe the particular methods used in the present study. Details on the tensor field represention a...

The problem of estimating depth information from two or more images of a scene is one which has received considerable attention over the years and a wide variety of methods have been proposed to solve it [Barnard and Fichsler, 1982; Fleck, 1991]. Methods based on correlation and methods using some form of feature matching between the images have found most widespread use. Of these, the latter have attracted increasing attention since the work of Marr [Marr, 1982], in which the features are zero-crossings on varying scales. These methods share an underlying basis of spatial domain operations.

In recent years, however, increasing interest has been shown in computational models of vision based primarily on a localized frequency domain representation - the Gabor representation [Gabor, 1946; Adelson and Bergen, 1985], first suggested in the context of computer vision by Granlund [Granlund, 1978].

The Application Visualization System software from Advanced Visual Systems Inc is an interactive visualization environment for scientists, engineers and technical professionals. This report contains a short overview of the AVS software packages and a discussion about its general performance. The software package has actively been used at the Computer Vision Laboratory, Linköping University, during the last three years. The AVS package has been used in many applications. Examples are generating images from a virtual environment, simulation of a controllable robot with a stereo camera head and visualization of multidimensional data structures. Lately we also have used AVS for handling communication between different processes which may be distributed on different machines. AVS was primarily developed as a tool for visualization of complex data sets. However, another important aspect of the software is that it can be used as an advanced workbench for controlling networks of Unix processes (including external ones on different machine types) using simple visual programming.

Adopting contrastive image-text pretrained models like CLIP towards video classification has gained attention due to its cost-effectiveness and competitive performance. However, recent works in this area face a trade-off. Fine-tuning the pretrained model to achieve strong supervised performance results in low zero-shot generalization. Similarly, freezing the backbone to retain zero-shot capability causes significant drop in supervised accuracy. Because of this, recent works in literature typically train separate models for supervised and zero-shot action recognition. In this work, we propose a multimodal prompt learning scheme that works to balance the supervised and zero-shot performance under a single unified training. Our prompting approach on the vision side caters for three aspects: 1) Global video-level prompts to model the data distribution; 2) Local frame-level prompts to provide per-frame discriminative conditioning; and 3) a summary prompt to extract a condensed video representation. Additionally, we define a prompting scheme on the text side to augment the textual context. Through this prompting scheme, we can achieve state-of-the-art zero-shot performance on Kinetics-600, HMDB51 and UCF101 while remaining competitive in the supervised setting. By keeping the pretrained backbone frozen, we optimize a much lower number of parameters and retain the existing general representation which helps achieve the strong zero-shot performance. Our codes/models will be released at https://github.com/TalalWasim/Vita-CLIP..

Pre-trained vision-language (V-L) models such as CLIP have shown excellent generalization ability to downstream tasks. However, they are sensitive to the choice of input text prompts and require careful selection of prompt templates to perform well. Inspired by the Natural Language Processing (NLP) literature, recent CLIP adaptation approaches learn prompts as the textual inputs to ne-tune CLIP for downstream tasks. We note that using prompting to adapt representations in a single branch of CLIP (language or vision) is sub-optimal since it does not allow the exibility to dynamically adjust both representation spaces on a downstream task. In this work, we propose Multi-modal Prompt Learning (MaPLe) for both vision and language branches to improve alignment between the vision and language representations. Our design promotes strong coupling between the vision-language prompts to ensure mutual synergy and discourages learning independent uni-modal solutions. Further, we learn separate prompts across different early stages to progressively model the stage-wise feature relationships to allow rich context learning. We evaluate the effectiveness of our approach on three representative tasks of generalization to novel classes, new target datasets and unseen domain shifts. Compared with the state-of-the-art method Co-CoOp, MaPLe exhibits favorable performance and achieves an absolute gain of 3.45% on novel classes and 2.72% on overall harmonic-mean, averaged over 11 diverse image recognition datasets. Our code and pre-trained models are available at https://github.com/muzairkhattak/multimodal-prompt-learning.

Deep neural networks (DNNs) have enabled astounding progress in several vision-based problems. Despite showing high predictive accuracy, recently, several works have revealed that they tend to provide overconfident predictions and thus are poorly calibrated. The majority of the works addressing the miscalibration of DNNs fall under the scope of classification and consider only in-domain predictions. However, there is little to no progress in studying the calibration of DNN-based object detection models, which are central to many vision-based safety-critical applications. In this paper, inspired by the train-time calibration methods, we propose a novel auxiliary loss formulation that explicitly aims to align the class confidence of bounding boxes with the accurateness of predictions (i.e. precision). Since the original formulation of our loss depends on the counts of true positives and false positives in a mini-batch, we develop a differentiable proxy of our loss that can be used during training with other application-specific loss functions. We perform extensive experiments on challenging in-domain and out-domain scenarios with six benchmark datasets including MS-COCO, Cityscapes, Sim10k, and BDD100k. Our results reveal that our train-time loss surpasses strong calibration baselines in reducing calibration error for both in and out-domain scenarios. Our source code and pre-trained models are available at https://github.com/akhtarvision/bpc_calibration

Although existing semi-supervised learning models achieve remarkable success in learning with unannotated in-distribution data, they mostly fail to learn on unlabeled data sampled from novel semantic classes due to their closed-set assumption. In this work, we target a pragmatic but under-explored Generalized Novel Category Discovery (GNCD) setting. The GNCD setting aims to categorize unlabeled training data coming from known and novel classes by leveraging the information of partially labeled known classes. We propose a two-stage Contrastive Affinity Learning method with auxiliary visual Prompts, dubbed PromptCAL, to address this challenging problem. Our approach discovers reliable pairwise sample affinities to learn better semantic clustering of both known and novel classes for the class token and visual prompts. First, we propose a discriminative prompt regularization loss to reinforce semantic discriminativeness of prompt-adapted pre-trained vision transformer for refined affinity relationships. Besides, we propose contrastive affinity learning to calibrate semantic representations based on our iterative semi-supervised affinity graph generation method for semantically-enhanced supervision. Extensive experimental evaluation demonstrates that our PromptCAL method is more effective in discovering novel classes even with limited annotations and surpasses the current state-of-the-art on generic and fine-grained benchmarks (e.g., with nearly 11% gain on CUB-200, and 9% on ImageNet-100) on overall accuracy. Our code is available at https: // github.com/ sheng- eatamath / PromptCAL.

Most previous co-salient object detection works mainly focus on extracting co-salient cues via mining the consistency relations across images while ignore explicit exploration of background regions. In this paper, we propose a Discriminative co-saliency and background Mining Transformer framework (DMT) based on several economical multi-grained correlation modules to explicitly mine both co-saliency and background information and effectively model their discrimination. Specifically, we first propose a region-to-region correlation module for introducing inter-image relations to pixel-wise segmentation features while maintaining computational efficiency. Then, we use two types of pre-defined tokens to mine co-saliency and background information via our proposed contrast-induced pixel-to-token correlation and co-saliency token-to-token correlation modules. We also design a token-guided feature refinement module to enhance the discriminability of the segmentation features under the guidance of the learned tokens. We perform iterative mutual promotion for the segmentation feature extraction and token construction. Experimental results on three benchmark datasets demonstrate the effectiveness of our proposed method. The source code is available at: https://github.com/dragonlee258079/DMT.

The pose-guided person image generation task requires synthesizing photorealistic images of humans in arbitrary poses. The existing approaches use generative adversarial networks that do not necessarily maintain realistic textures or need dense correspondences that struggle to handle complex deformations and severe occlusions. In this work, we show how denoising diffusion models can be applied for high-fidelity person image synthesis with strong sample diversity and enhanced mode coverage of the learnt data distribution. Our proposed Person Image Diffusion Model (PIDM) disintegrates the complex transfer problem into a series of simpler forward-backward denoising steps. This helps in learning plausible source-to-target transformation trajectories that result in faithful textures and undistorted appearance details. We introduce a texture diffusion module based on cross-attention to accurately model the correspondences between appearance and pose information available in source and target images. Further, we propose disentangled classifier-free guidance to ensure close resemblance between the conditional inputs and the synthesized output in terms of both pose and appearance information. Our extensive results on two large-scale benchmarks and a user study demonstrate the photorealism of our proposed approach under challenging scenarios. We also show how our generated images can help in downstream tasks. Code is available at https://github.com/ankanbhunia/PIDM.

We propose to leverage optical flow features for higher generalization power in semi-supervised video object segmentation. Optical flow is usually exploited as additional guidance information in many computer vision tasks. However, its relevance in video object segmentation was mainly in unsupervised settings or using the optical flow to warp or refine the previously predicted masks. Different from the latter, we propose to directly leverage the optical flow features in the target representation. We show that this enriched representation improves the encoder-decoder approach to the segmentation task. A model to extract the combined information from the optical flow and the image is proposed, which is then used as input to the target model and the decoder network. Unlike previous methods, e.g. in tracking where concatenation is used to integrate information from image data and optical flow, a simple yet effective attention mechanism is exploited in our work. Experiments on DAVIS 2017 and YouTube-VOS 2019 show that integrating the information extracted from optical flow into the original image branch results in a strong performance gain, especially in unseen classes which demonstrates its higher generalization power.

Traditionally, monocular 3D human pose estimation employs a machine learning model to predict the most likely 3D pose for a given input image. However, a single image can be highly ambiguous and induces multiple plausible solutions for the 2D-3D lifting step, which results in overly confident 3D pose predictors. To this end, we propose DiffPose, a conditional diffusion model that predicts multiple hypotheses for a given input image. Compared to similar approaches, our diffusion model is straightforward and avoids intensive hyperparameter tuning, complex network structures, mode collapse, and unstable training. Moreover, we tackle the problem of over-simplification of the intermediate representation of the common two-step approaches which first estimate a distribution of 2D joint locations via joint-wise heatmaps and consecutively use their maximum argument for the 3D pose estimation step. Since such a simplification of the heatmaps removes valid information about possibly correct, though labeled unlikely, joint locations, we propose to represent the heatmaps as a set of 2D joint candidate samples. To extract information about the original distribution from these samples, we introduce our embedding transformer which conditions the diffusion model. Experimentally, we show that DiffPose improves upon the state of the art for multi-hypothesis pose estimation by 3-5% for simple poses and outperforms it by a large margin for highly ambiguous poses.

A challenge in image based metrology and forensics is intrinsic camera calibration when the used camera is unavailable. The unavailability raises two questions. The first question is how to find the projection model that describes the camera, and the second is to detect incorrect models. In this work, we use off-the-shelf extended PnP-methods to find the model from 2D-3D correspondences, and propose a method for model validation. The most common strategy for evaluating a projection model is comparing different models’ residual variances—however, this naive strategy cannot distinguish whether the projection model is potentially underfitted or overfitted. To this end, we model the residual errors for each correspondence, individually scale all residuals using a predicted variance and test if the new residuals are drawn from a standard normal distribution. We demonstrate the effectiveness of our proposed validation in experiments on synthetic data, simulating 2D detection and Lidar measurements. Additionally, we provide experiments using data from an actual scene and compare non-camera access and camera access calibrations. Last, we use our method to validate annotations in MegaDepth.

Industrial defect detection is commonly addressed with anomaly detection (AD) methods where no or only incomplete data of potentially occurring defects is available. This work discovers previously unknown problems of studentteacher approaches for AD and proposes a solution, where two neural networks are trained to produce the same output for the defect-free training examples. The core assumption of student-teacher networks is that the distance between the outputs of both networks is larger for anomalies since they are absent in training. However, previous methods suffer from the similarity of student and teacher architecture, such that the distance is undesirably small for anomalies. For this reason, we propose asymmetric student-teacher networks (AST). We train a normalizing flow for density estimation as a teacher and a conventional feed-forward network as a student to trigger large distances for anomalies: The bijectivity of the normalizing flow enforces a divergence of teacher outputs for anomalies compared to normal data. Outside the training distribution the student cannot imitate this divergence due to its fundamentally different architecture. Our AST network compensates for wrongly estimated likelihoods by a normalizing flow, which was alternatively used for anomaly detection in previous work. We show that our method produces state-of-the-art results on the two currently most relevant defect detection datasets MVTec AD and MVTec 3D-AD regarding image-level anomaly detection on RGB and 3D data.

Class-Incremental Learning is a challenging problem in machine learning that aims to extend previously trained neural networks with new classes. This is especially useful if the system is able to classify new objects despite the original training data being unavailable. Although the semantic segmentation problem has received less attention than classification, it poses distinct problems and challenges, since previous and future target classes can be unlabeled in the images of a single increment. In this case, the background, past and future classes are correlated and there exists a background-shift.

In this paper, we address the problem of how to model unlabeled classes while avoiding spurious feature clustering of future uncorrelated classes. We propose to use Evidential Deep Learning to model the evidence of the classes as a Dirichlet distribution. Our method factorizes the problem into a separate foreground class probability, calculated by the expected value of the Dirichlet distribution, and an unknown class (background) probability corresponding to the uncertainty of the estimate. In our novel formulation, the background probability is implicitly modeled, avoiding the feature space clustering that comes from forcing the model to output a high background score for pixels that are not labeled as objects. Experiments on the incremental Pascal VOC and ADE20k benchmarks show that our method is superior to the state of the art, especially when repeatedly learning new classes with increasing number of increments.

Creative sketching or doodling is an expressive activity, where imaginative and previously unseen depictions of everyday visual objects are drawn. Creative sketch image generation is a challenging vision problem, where the task is to generate diverse, yet realistic creative sketches possessing the unseen composition of the visual-world objects. Here, we propose a novel coarse-to-fine two-stage framework, DoodleFormer, that decomposes the creative sketch generation problem into the creation of coarse sketch composition followed by the incorporation of fine-details in the sketch. We introduce graph-aware transformer encoders that effectively capture global dynamic as well as local static structural relations among different body parts. To ensure diversity of the generated creative sketches, we introduce a probabilistic coarse sketch decoder that explicitly models the variations of each sketch body part to be drawn. Experiments are performed on two creative sketch datasets: Creative Birds and Creative Creatures. Our qualitative, quantitative and human-based evaluations show that DoodleFormer outperforms the state-of-the-art on both datasets, yielding realistic and diverse creative sketches. On Creative Creatures, DoodleFormer achieves an absolute gain of 25 in Frechet inception distance (FID) over state-of-the-art. We also demonstrate the effectiveness of DoodleFormer for related applications of text to creative sketch generation, sketch completion and house layout generation. Code is available at: https://github.com/ ankanbhunia/doodleformer.

State-of-the-art transformer-based video instance segmentation (VIS) approaches typically utilize either single-scale spatio-temporal features or per-frame multi-scale features during the attention computations. We argue that such an attention computation ignores the multiscale spatio-temporal feature relationships that are crucial to tackle target appearance deformations in videos. To address this issue, we propose a transformer-based VIS framework, named MS-STS VIS, that comprises a novel multi-scale spatio-temporal split (MS-STS) attention module in the encoder. The proposed MS-STS module effectively captures spatio-temporal feature relationships at multiple scales across frames in a video. We further introduce an attention block in the decoder to enhance the temporal consistency of the detected instances in different frames of a video. Moreover, an auxiliary discriminator is introduced during training to ensure better foreground-background separability within the multiscale spatio-temporal feature space. We conduct extensive experiments on two benchmarks: Youtube-VIS (2019 and 2021). Our MS-STS VIS achieves state-of-the-art performance on both benchmarks. When using the ResNet50 backbone, our MS-STS achieves a mask AP of 50.1%, outperforming the best reported results in literature by 2.7% and by 4.8% at higher overlap threshold of AP75, while being comparable in model size and speed on Youtube-VIS 2019 val. set. When using the Swin Transformer backbone, MS-STS VIS achieves mask AP of 61.0% on Youtube-VIS 2019 val. set.

What constitutes an object? This has been a long-standing question in computer vision. Towards this goal, numerous learning-free and learning-based approaches have been developed to score objectness. However, they generally do not scale well across new domains and novel objects. In this paper, we advocate that existing methods lack a top-down supervision signal governed by human-understandable semantics. For the first time in literature, we demonstrate that Multi-modal Vision Transformers (MViT) trained with aligned image-text pairs can effectively bridge this gap. Our extensive experiments across various domains and novel objects show the state-of-the-art performance of MViTs to localize generic objects in images. Based on the observation that existing MViTs do not include multi-scale feature processing and usually require longer training schedules, we develop an efficient MViT architecture using multi-scale deformable attention and late vision-language fusion. We show the significance of MViT proposals in a diverse range of applications including open-world object detection, salient and camouflage object detection, supervised and self-supervised detection tasks. Further, MViTs can adaptively generate proposals given a specific language query and thus offer enhanced interactability.

Road networks are the core infrastructure for connected and autonomous vehicles, but creating meaningful representations for machine learning applications is a challenging task. In this work, we propose to integrate remote sensing vision data into road network data for improved embeddings with graph neural networks. We present a segmentation of road edges based on spatio-temporal road and traffic characteristics, which allows enriching the attribute set of road networks with visual features of satellite imagery and digital surface models. We show that both, the segmentation and the integration of vision data can increase performance on a road type classification task, and we achieve state-of-the-art performance on the OSM+DiDi Chuxing dataset on Chengdu, China.

Detecting abnormal events in video is commonly framed as a one-class classification task, where training videos contain only normal events, while test videos encompass both normal and abnormal events. In this scenario, anomaly detection is an open-set problem. However, some studies assimilate anomaly detection to action recognition. This is a closed-set scenario that fails to test the capability of systems at detecting new anomaly types. To this end, we propose UBnormal, a new supervised open-set benchmark composed of multiple virtual scenes for video anomaly detection. Unlike existing data sets, we introduce abnormal events annotated at the pixel level at training time, for the first time enabling the use of fully-supervised learning methods for abnormal event detection. To preserve the typical open-set formulation, we make sure to include dis-joint sets of anomaly types in our training and test collections of videos. To our knowledge, UBnormal is the first video anomaly detection benchmark to allow a fair head-to-head comparison between one-class open-set models and supervised closed-set models, as shown in our experiments. Moreover, we provide empirical evidence showing that UB-normal can enhance the performance of a state-of-the-art anomaly detection framework on two prominent data sets, Avenue and ShanghaiTech. Our benchmark is freely available at https://github.com/lilygeorgescu/UBnormal.

Anomaly detection is commonly pursued as a one-class classification problem, where models can only learn from normal training samples, while being evaluated on both normal and abnormal test samples. Among the successful approaches for anomaly detection, a distinguished category of methods relies on predicting masked information (e.g. patches, future frames, etc.) and leveraging the reconstruction error with respect to the masked information as an abnormality score. Different from related methods, we propose to integrate the reconstruction-based functionality into a novel self-supervised predictive architectural building block. The proposed self-supervised block is generic and can easily be incorporated into various state-of-the-art anomaly detection methods. Our block starts with a convolutional layer with dilated filters, where the center area of the receptive field is masked. The resulting activation maps are passed through a channel attention module. Our block is equipped with a loss that minimizes the reconstruction error with respect to the masked area in the receptive field. We demonstrate the generality of our block by integrating it into several state-of-the-art frameworks for anomaly detection on image and video, providing empirical evidence that shows considerable performance improvements on MVTec AD, Avenue, and ShanghaiTech. We release our code as open source at: https://github.com/ ristea/sspcab.

Deep learning models tend to forget their earlier knowledge while incrementally learning new tasks. This behavior emerges because the parameter updates optimized for the new tasks may not align well with the updates suitable for older tasks. The resulting latent representation mismatch causes forgetting. In this work, we propose ELI: Energy-based Latent Aligner for Incremental Learning, which first learns an energy manifold for the latent representations such that previous task latents will have low energy and the current task latents have high energy values. This learned manifold is used to counter the representational shift that happens during incremental learning. The implicit regularization that is offered by our proposed methodology can be used as a plug-and-play module in existing incremental learning methodologies. We validate this through extensive evaluation on CIFAR-100, ImageNet subset, ImageNet 1k and Pascal VOC datasets. We observe consistent improvement when ELI is added to three prominent methodologies in class-incremental learning, across multiple incremental settings. Further, when added to the state-of-the-art incremental object detector, ELI provides over 5% improvement in detection accuracy, corroborating its effectiveness and complementary advantage to the existing art. Code is available at: https://github.com/JosephKJ/ELI.

Modern handheld devices can acquire burst image sequence in a quick succession. However, the individual acquired frames suffer from multiple degradations and are misaligned due to camera shake and object motions. The goal of Burst Image Restoration is to effectively combine complimentary cues across multiple burst frames to generate high-quality outputs. Towards this goal, we develop a novel approach by solely focusing on the effective information exchange between burst frames, such that the degradations get filtered out while the actual scene details are preserved and enhanced. Our central idea is to create a set of pseudo-burst features that combine complimentary information from all the input burst frames to seamlessly exchange information. However, the pseudo-burst cannot be successfully created unless the individual burst frames are properly aligned to discount interframe movements. Therefore, our approach initially extracts pre-processed features from each burst frame and matches them using an edge-boosting burst alignment module. The pseudo-burst features are then created and enriched using multi-scale contextual information. Our final step is to adaptively aggregate information from the pseudo-burst features to progressively increase resolution in multiple stages while merging the pseudo-burst features. In comparison to existing works that usually follow a late fusion scheme with single-stage upsampling, our approach performs favorably, delivering state-of-the-art performance on burst super-resolution, burst low-light image enhancement and burst denoising tasks. The source code and pre-trained models are available at https://github.com/akshaydudhane16/BIPNet.

Modern digital cameras generally count on image signal processing (ISP) pipelines for producing naturalistic RGB images. Nevertheless, in comparison to DSLR cameras, low-quality images are generally output from portable mobile devices due to their physical limitations. The synthesized low-quality images usually have multiple degradations - low-resolution owing to small camera sensors, mosaic patterns on account of camera filter array and subpixel shifts due to camera motion. Such degradation usually restrain the performance of single image super-resolution methodologies for retrieving high-resolution (HR) image from a single low-resolution (LR) image. Burst image super-resolution aims at restoring a photo-realistic HR image by capturing the abundant information from multiple LR images. Lately, the soaring popularity of burst photography has made multi-frame processing an attractive solution for overcoming the limitations of single image processing. In our work, we thus aim to propose a generic architecture, adaptive feature consolidation network (AFCNet) for multi-frame processing. To alleviate the challenge of effectively modelling the long-range dependency problem, that multi-frame approaches struggle to solve, we utilize encoder-decoder based transformer backbone which learns multi-scale local-global representations. We propose feature alignment module to align LR burst frame features. Further, the aligned features are fused and reconstructed by abridged pseudo-burst fusion module and adaptive group upsampling modules, respectively. Our proposed approach clearly outperforms the other existing state-of-the-art techniques on benchmark datasets. The experimental results illustrate the effectiveness and generality of our proposed framework in upgrading the visual quality of HR images.

In this paper, we argue that modern pre-integration methods for inertial measurement units (IMUs) are accurate enough to ignore the drift for short time intervals. This allows us to consider a simplified camera model, which in turn admits further intrinsic calibration. We develop the first-ever solver to jointly solve the relative pose problem with unknown and equal focal length and radial distortion profile while utilizing the IMU data. Furthermore, we show significant speed-up compared to state-of-the-art algorithms, with small or negligible loss in accuracy for partially calibrated setups.The proposed algorithms are tested on both synthetic and real data, where the latter is focused on navigation using unmanned aerial vehicles (UAVs). We evaluate the proposed solvers on different commercially available low-cost UAVs, and demonstrate that the novel assumption on IMU drift is feasible in real-life applications. The extended intrinsic auto-calibration enables us to use distorted input images, making tedious calibration processes obsolete, compared to current state-of-the-art methods. Code available at: https://github.com/marcusvaltonen/DronePoseLib.

Emerging from low-level vision theory, steerable filters found their counterpart in prior work on steerable convolutional neural networks equivariant to rigid transformations. In our work, we propose a steerable feed-forward learning-based approach that consists of neurons with spherical decision surfaces and operates on point clouds. Such spherical neurons are obtained by conformal embedding of Euclidean space and have recently been revisited in the context of learning representations of point sets. Focusing on 3D geometry, we exploit the isometry property of spherical neurons and derive a 3D steerability constraint. After training spherical neurons to classify point clouds in a canonical orientation, we use a tetrahedron basis to quadruplicate the neurons and construct rotation-equivariant spherical filter banks. We then apply the derived constraint to interpolate the filter bank outputs and, thus, obtain a rotation-invariant network. Finally, we use a synthetic point set and real-world 3D skeleton data to verify our theoretical findings. The code is available at https://github.com/pavlo-melnyk/steerable-3d-neurons.

This article proposes an architecture, which allows the prediction of intention by internally simulating perceptual states represented by action pattern vectors. To this end, associative self-organising neural networks (A-SOM) is utilised to build a hierarchical cognitive architecture for recognition and simulation of the skeleton based human actions. The abilities of the proposed architecture in recognising and predicting actions is evaluated in experiments using three different datasets of 3D actions. Based on the experiments of this article, applying internally simulated perceptual states represented by action pattern vectors improves the performance of the recognition task in all experiments. Furthermore, internal simulation of perception addresses the problem of having limited access to the sensory input, and also the future prediction of the consecutive perceptual sequences. The performance of the system is compared and discussed with similar architecture using self-organizing neural networks (SOM).

While the untargeted black-box transferability of adversarial perturbations has been extensively studied before, changing an unseen models decisions to a specific targeted class remains a challenging feat. In this paper, we propose a new generative approach for highly transferable targeted perturbations (TTP). We note that the existing methods are less suitable for this task due to their reliance on class-boundary information that changes from one model to another, thus reducing transferability. In contrast, our approach matches the perturbed image distribution with that of the target class, leading to high targeted transferability rates. To this end, we propose a new objective function that not only aligns the global distributions of source and target images, but also matches the local neighbourhood structure between the two domains. Based on the proposed objective, we train a generator function that can adaptively synthesize perturbations specific to a given input. Our generative approach is independent of the source or target domain labels, while consistently performs well against state-of-the-art methods on a wide range of attack settings. As an example, we achieve 32.63% target transferability from (an adversarially weak) VGG19(BN) to (a strong) WideResNet on ImageNet val. set, which is 4x higher than the previous best generative attack and 16x better than instance-specific iterative attack. Code is available at: https://github.com/Muzammal-Naseer/TTP.

Deep neural networks have achieved remarkable performance on a range of classification tasks, with softmax cross-entropy (CE) loss emerging as the de-facto objective function. The CE loss encourages features of a class to have a higher projection score on the true class-vector compared to the negative classes. However, this is a relative constraint and does not explicitly force different class features to be well-separated. Motivated by the observation that ground-truth class representations in CE loss are orthogonal (one-hot encoded vectors), we develop a novel loss function termed `Orthogonal Projection Loss (OPL) which imposes orthogonality in the feature space. OPL augments the properties of CE loss and directly enforces inter-class separation alongside intra-class clustering in the feature space through orthogonality constraints on the mini-batch level. As compared to other alternatives of CE, OPL offers unique advantages e.g., no additional learnable parameters, does not require careful negative mining and is not sensitive to the batch size. Given the plug-and-play nature of OPL, we evaluate it on a diverse range of tasks including image recognition (CIFAR-100), large-scale classification (ImageNet), domain generalization (PACS) and few-shot learning (miniImageNet, CIFAR-FS, tiered-ImageNet and Meta-dataset) and demonstrate its effectiveness across the board. Furthermore, OPL offers better robustness against practical nuisances such as adversarial attacks and label noise. Code is available at: https://github.com/kahnchana/opl.

Multi-label zero-shot learning (ZSL) is a more realistic counter-part of standard single-label ZSL since several objects can co-exist in a natural image. However, the occurrence of multiple objects complicates the reasoning and requires region-specific processing of visual features to preserve their contextual cues. We note that the best existing multi-label ZSL method takes a shared approach towards attending to region features with a common set of attention maps for all the classes. Such shared maps lead to diffused attention, which does not discriminatively focus on relevant locations when the number of classes are large. Moreover, mapping spatially-pooled visual features to the class semantics leads to inter-class feature entanglement, thus hampering the classification. Here, we propose an alternate approach towards region-based discriminability-preserving multi-label zero-shot classification. Our approach maintains the spatial resolution to preserve region-level characteristics and utilizes a bi-level attention module (BiAM) to enrich the features by incorporating both region and scene context information. The enriched region-level features are then mapped to the class semantics and only their class predictions are spatially pooled to obtain image-level predictions, thereby keeping the multi-class features disentangled. Our approach sets a new state of the art on two large-scale multi-label zero-shot benchmarks: NUS-WIDE and Open Images. On NUS-WIDE, our approach achieves an absolute gain of 6.9% mAP for ZSL, compared to the best published results.

The Visual Object Tracking challenge VOT2021 is the ninth annual tracker benchmarking activity organized by the VOT initiative. Results of 71 trackers are presented; many are state-of-the-art trackers published at major computer vision conferences or in journals in recent years. The VOT2021 challenge was composed of four sub-challenges focusing on different tracking domains: (i) VOT-ST2021 challenge focused on short-term tracking in RGB, (ii) VOT-RT2021 challenge focused on "real-time" short-term tracking in RGB, (iii) VOT-LT2021 focused on long-term tracking, namely coping with target disappearance and reappearance and (iv) VOT-RGBD2021 challenge focused on long-term tracking in RGB and depth imagery. The VOT-ST2021 dataset was refreshed, while VOT-RGBD2021 introduces a training dataset and sequestered dataset for winner identification. The source code for most of the trackers, the datasets, the evaluation kit and the results along with the source code for most trackers are publicly available at the challenge website(1).

Humans have a natural instinct to identify unknown object instances in their environments. The intrinsic curiosity about these unknown instances aids in learning about them, when the corresponding knowledge is eventually available. This motivates us to propose a novel computer vision problem called: Open World Object Detection, where a model is tasked to: 1) identify objects that have not been introduced to it as unknown, without explicit supervision to do so, and 2) incrementally learn these identified unknown categories without forgetting previously learned classes, when the corresponding labels are progressively received. We formulate the problem, introduce a strong evaluation protocol and provide a novel solution, which we call ORE: Open World Object Detector, based on contrastive clustering and energy based unknown identification. Our experimental evaluation and ablation studies analyse the efficacy of ORE in achieving Open World objectives. As an interesting by-product, we find that identifying and characterising unknown instances helps to reduce confusion in an incremental object detection setting, where we achieve state-ofthe-art performance, with no extra methodological effort. We hope that our work will attract further research into this newly identified, yet crucial research direction.

Safe robot navigation in a dynamic environment, requires the trajectories of each independently moving object (IMO). We present the novel and effective system Sequential Hierarchical Ransac Estimation (Shire) designed for this purpose. The system uses a stereo camera stream to find the objects and trajectories in real time. Shire detects moving objects using geometric consistency and finds their trajectories using bundle adjustment. Relying on geometric consistency allows the system to handle objects regardless of semantic class, unlike approaches based on semantic segmentation. Most Visual Odometry (VO) systems are inherently limited to single motion by the choice of tracker. This limitation allows for efficient and robust ego-motion estimation in real time, but preclude tracking the multiple motions sought. Shire instead uses a generic tracker and achieves accurate VO and IMO estimates using track analysis. This removes the restriction to a single motion while retaining the real-time performance required for live navigation. We evaluate the system by bounding box intersection over union and ID persistence on a public dataset, collected from an autonomous test vehicle driving in real traffic. We also show the velocities of estimated IMOs. We investigate variations of the system that provide trade offs between accuracy, performance and limitations.

We investigate a novel deep-learning-based approach to estimate uncertainty in stereo disparity prediction networks. Current state-of-the-art methods often formulate disparity prediction as a regression problem with a single scalar output in each pixel. This can be problematic in practical applications as in many cases there might not exist a single well defined disparity, for example in cases of occlusions or at depth-boundaries. While current neural-network-based disparity estimation approaches  obtain good performance on benchmarks, the disparity prediction is treated as a black box at inference time. In this paper we show that by formulating the learning problem as a regression with a distribution target, we obtain a robust estimate of the uncertainty in each pixel, while maintaining the performance of the original method. The proposed method is evaluated both on a large-scale standard benchmark, as well on our own data. We also show that the uncertainty estimate significantly improves by maximizing the uncertainty in those pixels that have no well defined disparity during learning.

In this paper we present a novel algorithm for onboard radial distortion correction for unmanned aerial vehicles (UAVs) equipped with an inertial measurement unit (IMU), that runs in real-time. This approach makes calibration procedures redundant, thus allowing for exchange of optics extemporaneously. By utilizing the IMU data, the cameras can be aligned with the gravity direction. This allows us to work with fewer degrees of freedom, and opens up for further intrinsic calibration. We propose a fast and robust minimal solver for simultaneously estimating the focal length, radial distortion profile and motion parameters from homographies. The proposed solver is tested on both synthetic and real data, and perform better or on par with state-of-the-art methods relying on pre-calibration procedures. Code available at: https://github.com/marcusvaltonen/HomLib. 1

In this paper, we present a state-of-the-art reinforcement learning method for autonomous driving. Our approach employs temporal difference learning in a Bayesian framework to learn vehicle control signals from sensor data. The agent has access to images from a forward facing camera, which are pre-processed to generate semantic segmentation maps. We trained our system using both ground truth and estimated semantic segmentation input. Based on our observations from a large set of experiments, we conclude that training the system on ground truth input data leads to better performance than training the system on estimated input even if estimated input is used for evaluation. The system is trained and evaluated in a realistic simulated urban environment using the CARLA simulator. The simulator also contains a benchmark that allows for comparing to other systems and methods. The required training time of the system is shown to be lower and the performance on the benchmark superior to competing approaches.

Semi-supervised video object segmentation is a challenging task that aims to segment a target throughout a video sequence given an initial mask at the first frame. Discriminative approaches have demonstrated competitive performance on this task at a sensible complexity. These approaches typically formulate the problem as a one-versus-one classification between the target and the background. However, in reality, a video sequence usually encompasses a target, background, and possibly other distracting objects. Those objects increase the risk of introducing false positives, especially if they share visual similarities with the target. Therefore, it is more effective to separate distractors from the background, and handle them independently.

We propose a one-versus-many scheme to address this situation by separating distractors into their own class. This separation allows imposing special attention to challenging regions that are most likely to degrade the performance. We demonstrate the prominence of this formulation by modifying the learning-what-to-learn method to be distractor-aware. Our proposed approach sets a new state-of-the-art on the DAVIS val dataset, and improves over the baseline on the DAVIS test-dev benchmark by 4.8 percent points.

Unsupervised learning of anomaly detection in high-dimensional data, such as images, is a challenging problem recently subject to intense research. Through careful modelling of the data distribution of normal samples, it is possible to detect deviant samples, so called anomalies. Generative Adversarial Networks (GANs) can model the highly complex, high-dimensional data distribution of normal image samples, and have shown to be a suitable approach to the problem. Previously published GAN-based anomaly detection methods often assume that anomaly-free data is available for training. However, this assumption is not valid in most real-life scenarios, a.k.a. in the wild. In this work, we evaluate the effects of anomaly contaminations in the training data on state-of-the-art GAN-based anomaly detection methods. As expected, detection performance deteriorates. To address this performance drop, we propose to add an additional encoder network already at training time and show that joint generator-encoder training stratifies the latent space, mitigating the problem with contaminated data. We show experimentally that the norm of a query image in this stratified latent space becomes a highly significant cue to discriminate anomalies from normal data. The proposed method achieves state-of-the-art performance on CIFAR-10 as well as on a large, previously untested dataset with cell images.

Recent innovations in microelectronic and semiconductor technology enable the creation of smaller and economical hyperspectral cameras. A filter combined camera with advanced scanning module is a game changer that extends the application of miniature hyperspectral imagers to many security domains. This work presents the assessment of the imager L4 from Glana Sensors to detect concealed targets in woodland areas. Several target detection methods were applied to a collection of scenes acquired under various illumination conditions and containing different materials. The potential and limitations of this new imaging device in the context of difficult target detection in forested area are evaluated and discussed.

Video object segmentation (VOS) is a highly challengingproblem, since the target object is only defined by a first-frame refer-ence mask during inference. The problem of how to capture and utilizethis limited information to accurately segment the target remains a fun-damental research question. We address this by introducing an end-to-end trainable VOS architecture that integrates a differentiable few-shotlearner. Our learner is designed to predict a powerful parametric modelof the target by minimizing a segmentation error in the first frame. Wefurther go beyond the standard few-shot learning paradigm by learningwhat our target model should learn in order to maximize segmentationaccuracy. We perform extensive experiments on standard benchmarks.Our approach sets a new state-of-the-art on the large-scale YouTube-VOS 2018 dataset by achieving an overall score of 81.5, corresponding toa 2.6% relative improvement over the previous best result. The code andmodels are available at https://github.com/visionml/pytracking.

Adversarial examples can cause catastrophic mistakes in Deep Neural Network (DNNs) based vision systems e.g., for classification, segmentation and object detection. The vulnerability of DNNs against such attacks can prove a major roadblock towards their real-world deployment. Transferability of adversarial examples demand generalizable defenses that can provide cross-task protection. Adversarial training that enhances robustness by modifying target model’s parameters lacks such generalizability. On the other hand, different input processing based defenses fall short in the face of continuously evolving attacks. In this paper, we take the first step to combine the benefits of both approaches and propose a self-supervised adversarial training mechanism in the input space. By design, our defense is a generalizable approach and provides significant robustness against the unseen adversarial attacks (\eg by reducing the success rate of translation-invariant ensemble attack from 82.6% to 31.9% in comparison to previous state-of-the-art). It can be deployed as a plug-and-play solution to protect a variety of vision systems, as we demonstrate for the case of classification, segmentation and detection.

One of the attractive characteristics of deep neural networks is their ability to transfer knowledge obtained in one domain to other related domains. As a result, high-quality networks can be trained in domains with relatively little training data. This property has been extensively studied for discriminative networks but has received significantly less attention for generative models. Given the often enormous effort required to train GANs, both computationally as well as in the dataset collection, the re-use of pretrained GANs is a desirable objective. We propose a novel knowledge transfer method for generative models based on mining the knowledge that is most beneficial to a specific target domain, either from a single or multiple pretrained GANs. This is done using a miner network that identifies which part of the generative distribution of each pretrained GAN outputs samples closest to the target domain. Mining effectively steers GAN sampling towards suitable regions of the latent space, which facilitates the posterior finetuning and avoids pathologies of other methods such as mode collapse and lack of flexibility. We perform experiments on several complex datasets using various GAN architectures (BigGAN, Progressive GAN) and show that the proposed method, called MineGAN, effectively transfers knowledge to domains with few target images, outperforming existing methods. In addition, MineGAN can successfully transfer knowledge from multiple pretrained GANs. Our code is available at: https://github.com/yaxingwang/MineGAN.

Humans can continuously learn new knowledge as their experience grows. In contrast, previous learning in deep neural networks can quickly fade out when they are trained on a new task. In this paper, we hypothesize this problem can be avoided by learning a set of generalized parameters, that are neither specific to old nor new tasks. In this pursuit, we introduce a novel meta-learning approach that seeks to maintain an equilibrium between all the encountered tasks. This is ensured by a new meta-update rule which avoids catastrophic forgetting. In comparison to previous meta-learning techniques, our approach is task-agnostic. When presented with a continuum of data, our model automatically identifies the task and quickly adapts to it with just a single update. We perform extensive experiments on five datasets in a class-incremental setting, leading to significant improvements over the state of the art methods (e.g., a 21.3% boost on CIFAR100 with 10 incremental tasks). Specifically, on large-scale datasets that generally prove difficult cases for incremental learning, our approach delivers absolute gains as high as 19.1% and 7.4% on ImageNet and MS-Celeb datasets, respectively.

The Visual Object Tracking challenge VOT2019 is the seventh annual tracker benchmarking activity organized by the VOT initiative. Results of 81 trackers are presented; many are state-of-the-art trackers published at major computer vision conferences or in journals in the recent years. The evaluation included the standard VOT and other popular methodologies for short-term tracking analysis as well as the standard VOT methodology for long-term tracking analysis. The VOT2019 challenge was composed of five challenges focusing on different tracking domains: (i) VOT-ST2019 challenge focused on short-term tracking in RGB, (ii) VOT-RT2019 challenge focused on "real-time" short-term tracking in RGB, (iii) VOT-LT2019 focused on long-term tracking namely coping with target disappearance and reappearance. Two new challenges have been introduced: (iv) VOT-RGBT2019 challenge focused on short-term tracking in RGB and thermal imagery and (v) VOT-RGBD2019 challenge focused on long-term tracking in RGB and depth imagery. The VOT-ST2019, VOT-RT2019 and VOT-LT2019 datasets were refreshed while new datasets were introduced for VOT-RGBT2019 and VOT-RGBD2019. The VOT toolkit has been updated to support both standard short-term, long-term tracking and tracking with multi-channel imagery. Performance of the tested trackers typically by far exceeds standard baselines. The source code for most of the trackers is publicly available from the VOT page. The dataset, the evaluation kit and the results are publicly available at the challenge website(1).

Common object counting in a natural scene is a challenging problem in computer vision with numerous real-world applications. Existing image-level supervised common object counting approaches only predict the global object count and rely on additional instance-level supervision to also determine object locations. We propose an image-level supervised approach that provides both the global object count and the spatial distribution of object instances by constructing an object category density map. Motivated by psychological studies, we further reduce image-level supervision using a limited object count information (up to four). To the best of our knowledge, we are the first to propose image-level supervised density map estimation for common object counting and demonstrate its effectiveness in image-level supervised instance segmentation. Comprehensive experiments are performed on the PASCAL VOC and COCO datasets. Our approach outperforms existing methods, including those using instance-level supervision, on both datasets for common object counting. Moreover, our approach improves state-of-the-art image-level supervised instance segmentation [34] with a relative gain of 17.8% in terms of average best overlap, on the PASCAL VOC 2012 dataset.

Pedestrian detection relying on deep convolution neural networks has made significant progress. Though promising results have been achieved on standard pedestrians, the performance on heavily occluded pedestrians remains far from satisfactory. The main culprits are intra-class occlusions involving other pedestrians and inter-class occlusions caused by other objects, such as cars and bicycles. These result in a multitude of occlusion patterns. We propose an approach for occluded pedestrian detection with the following contributions. First, we introduce a novel mask-guided attention network that fits naturally into popular pedestrian detection pipelines. Our attention network emphasizes on visible pedestrian regions while suppressing the occluded ones by modulating full body features. Second, we empirically demonstrate that coarse-level segmentation annotations provide reasonable approximation to their dense pixel-wise counterparts. Experiments are performed on CityPersons and Caltech datasets. Our approach sets a new state-of-the-art on both datasets. Our approach obtains an absolute gain of 9.5% in log-average miss rate, compared to the best reported results [31] on the heavily occluded HO pedestrian set of CityPersons test set. Further, on the HO pedestrian set of Caltech dataset, our method achieves an absolute gain of 5.0% in log-average miss rate, compared to the best reported results [13]. Code and models are available at: https://github.com/Leotju/MGAN.

Adversarial examples reveal the blind spots of deep neural networks (DNNs) and represent a major concern for security-critical applications. The transferability of adversarial examples makes real-world attacks possible in black-box settings, where the attacker is forbidden to access the internal parameters of the model. The underlying assumption in most adversary generation methods, whether learning an instance-specific or an instance-agnostic perturbation, is the direct or indirect reliance on the original domain-specific data distribution. In this work, for the first time, we demonstrate the existence of domain-invariant adversaries, thereby showing common adversarial space among different datasets and models. To this end, we propose a framework capable of launching highly transferable attacks that crafts adversarial patterns to mislead networks trained on entirely different domains. For instance, an adversarial function learned on Paintings, Cartoons or Medical images can successfully perturb ImageNet samples to fool the classifier, with success rates as high as similar to 99% (l(infinity) &lt;= 10). The core of our proposed adversarial function is a generative network that is trained using a relativistic supervisory signal that enables domain-invariant perturbations. Our approach sets the new state-of-the-art for fooling rates, both under the white-box and black-box scenarios. Furthermore, despite being an instance-agnostic perturbation function, our attack outperforms the conventionally much stronger instance-specific attack methods.

While recent years have witnessed astonishing improvements in visual tracking robustness, the advancements in tracking accuracy have been limited. As the focus has been directed towards the development of powerful classifiers, the problem of accurate target state estimation has been largely overlooked. In fact, most trackers resort to a simple multi-scale search in order to estimate the target bounding box. We argue that this approach is fundamentally limited since target estimation is a complex task, requiring highlevel knowledge about the object. We address this problem by proposing a novel tracking architecture, consisting of dedicated target estimation and classification components. High level knowledge is incorporated into the target estimation through extensive offline learning. Our target estimation component is trained to predict the overlap between the target object and an estimated bounding box. By carefully integrating targetspecific information, our approach achieves previously unseen bounding box accuracy. We further introduce a classification component that is trained online to guarantee high discriminative power in the presence of distractors. Our final tracking framework sets a new state-of-the-art on five challenging benchmarks. On the new large-scale TrackingNet dataset, our tracker ATOM achieves a relative gain of 15% over the previous best approach, while running at over 30 FPS. Code and models are available at https://github.com/visionml/pytracking.

Deep learning requires large amounts of annotated data. Manual annotation of objects in video is, regardless of annotation type, a tedious and time-consuming process. In particular, for scarcely used image modalities human annotationis hard to justify. In such cases, semi-automatic annotation provides an acceptable option.

In this work, a recursive, semi-automatic annotation method for video is presented. The proposed method utilizesa state-of-the-art video object segmentation method to propose initial annotations for all frames in a video based on only a few manual object segmentations. In the case of a multi-modal dataset, the multi-modality is exploited to refine the proposed annotations even further. The final tentative annotations are presented to the user for manual correction.

The method is evaluated on a subset of the RGBT-234 visual-thermal dataset reducing the workload for a human annotator with approximately 78% compared to full manual annotation. Utilizing the proposed pipeline, sequences are annotated for the VOT-RGBT 2019 challenge.

Trackers based on discriminative correlation filters (DCF) have recently seen widespread success and in this work we dive into their numerical core. DCF-based trackers interleave learning of the target detector and target state inference based on this detector. Whereas the original formulation includes a closed-form solution for the filter learning, recently introduced improvements to the framework no longer have known closed-form solutions. Instead a large-scale linear least squares problem must be solved each time the detector is updated. We analyze the procedure used to optimize the detector and let the popular scheme introduced with ECO serve as a baseline. The ECO implementation is revisited in detail and several mechanisms are provided with alternatives. With comprehensive experiments we show which configurations are superior in terms of tracking capabilities and optimization performance.

Availability of large training datasets was essential for the recent advancement and success of deep learning methods. Due to the difficulties related to biometric data collection, datasets with age and gender annotations are scarce and usually limited in terms of size and sample diversity. Web-scraping approaches for automatic data collection can produce large amounts weakly labeled noisy data. The unsupervised facial biometric data filtering method presented in this paper greatly reduces label noise levels in web-scraped facial biometric data. Experiments on two large state-of-the-art web-scraped facial datasets demonstrate the effectiveness of the proposed method, with respect to training and validation scores, training convergence, and generalization capabilities of trained age and gender estimators.

We present Lambda Twist; a novel P3P solver which is accurate, fast and robust. Current state-of-the-art P3P solvers find all roots to a quartic and discard geometrically invalid and duplicate solutions in a post-processing step. Instead of solving a quartic, the proposed P3P solver exploits the underlying elliptic equations which can be solved by a fast and numerically accurate diagonalization. This diagonalization requires a single real root of a cubic which is then used to find the, up to four, P3P solutions. Unlike the direct quartic solvers our method never computes geometrically invalid or duplicate solutions.

Extensive evaluation on synthetic data shows that the new solver has better numerical accuracy and is faster compared to the state-of-the-art P3P implementations. Implementation and benchmark are available on github.