The deployment of large, black-box foundation models for medical image classification is often hindered by the high cost of acquiring large, task-specific labeled datasets for fine-tuning. While active learning (AL) presents a promising solution, many state-of-the-art AL methods are computationally expensive or require full access to internal model parameters. We present VALIANT (Visual Adaptation and Learning Integration for Active learNing Tasks), a new active learning framework designed to efficiently adapt black-box foundation models by overcoming these limitations. VALIANT introduces a lightweight Visual Prompt Decoder (VIPD), trained via unsupervised Zero-Order Optimization (ZOO), to generate task-specific visual prompts without internal model access. Our core contribution is a perturbation-based ranking strategy that leverages this VIPD to formulate a computationally efficient, gradient-aware informativeness metric. This metric, which we term prompt instability, identifies the most impactful samples for the labeling budget. VALIANT further enhances this process by incorporating anatomical information from unsupervised segmentation maps to generate more discriminative visual prompts. Extensive evaluations on multiple medical datasets demonstrate VALIANT’s superior performance and significant reduction in labeling costs compared to a range of existing active learning techniques, positioning it as a scalable and practical solution for medical image analysis.

Zero-shot classification capabilities naturally arise in models trained within a vision-language contrastive framework. Despite their classification prowess, these models struggle in dense tasks like zero-shot open-vocabulary segmentation. This deficiency is often attributed to the absence of localization cues in captions and the intertwined nature of the learning process, which encompasses both image/text representation learning and cross-modality alignment. To tackle these issues, we propose SimZSS, a $\textbf{Sim}$ple framework for open-vocabulary $\textbf{Z}$ero-$\textbf{S}$hot $\textbf{S}$egmentation. The method is founded on two key principles: i) leveraging frozen vision-only models that exhibit spatial awareness while exclusively aligning the text encoder and ii) exploiting the discrete nature of text and linguistic knowledge to pinpoint local concepts within captions. By capitalizing on the quality of the visual representations, our method requires only image-caption pair datasets and adapts to both small curated and large-scale noisy datasets. When trained on COCO Captions across 8 GPUs, SimZSS achieves state-of-the-art results on 7 out of 8 benchmark datasets in less than 15 minutes. Our code and pretrained models are publicly available at https://github.com/tileb1/simzss.

Despite the impressive progress of recent pretraining methods on multimodal tasks, existing methods are inherently biased towards either spatial modeling (e.g., CLIP) or temporal modeling (e.g., V-JEPA), limiting their joint capture of spatial details and temporal dynamics. To this end, we propose UniViT, a cluster-driven unified self-supervised learning framework that effectively captures the structured semantics of both image spatial content and video temporal dynamics through event-level and object-level clustering and discrimination. Specifically, we leverage offline clustering to generate semantic clusters across both modalities. For videos, multi-granularity event-level clustering progressively expands from single-event to structured multi-event segments, capturing coarse-to-fine temporal semantics; for images, object-level clustering captures fine-grained spatial semantics. However, while global clustering provides semantically consistent clusters, it lacks modeling of structured semantic relations (e.g., temporal event structures). To address this, we introduce a contrastive objective that leverages these semantic clusters as pseudo-label supervision to explicitly enforce structural constraints, including temporal event relations and spatial object co-occurrences, capturing structured semantics beyond categories. Meanwhile, UniViT jointly embeds structured object-level and event-level semantics into a unified representation space. Furthermore, UniViT introduces two key components: (i) Unified Rotary Position Embedding integrates relative positional embedding with frequency-aware dimension allocation to support position-invariant semantic learning and enhance the stability of structured semantics in the discrimination stage; and (ii) Variable Spatiotemporal Streams adapt to inputs of varying frame lengths, addressing the rigidity of conventional fixed-input approaches. Extensive experiments across varying model scales demonstrate that UniViT achieves state-of-the-art performance on linear probing, attentive probing, question answering, and spatial understanding tasks.

This paper introduces **ReservoirTTA**, a novel plug–in framework designed for prolonged test–time adaptation (TTA) in scenarios where the test domain continuously shifts over time, including cases where domains recur or evolve gradually. At its core, ReservoirTTA maintains a reservoir of domain-specialized models—an adaptive test-time model ensemble—that both detects new domains via online clustering over style features of incoming samples and routes each sample to the appropriate specialized model, and thereby enables domain-specific adaptation. This multi-model strategy overcomes key limitations of single model adaptation, such as catastrophic forgetting, inter-domain interference, and error accumulation, ensuring robust and stable performance on sustained non-stationary test distributions. Our theoretical analysis reveals key components that bound parameter variance and prevent model collapse, while our plug–in TTA module mitigates catastrophic forgetting of previously encountered domains. Extensive experiments on scene-level corruption benchmarks (ImageNet-C, CIFAR-10/100-C), object-level style shifts (DomainNet-126, PACS), and semantic segmentation (Cityscapes→ACDC) — covering recurring and continuously evolving domain shifts — show that ReservoirTTA substantially improves adaptation accuracy and maintains stable performance across prolonged, recurring shifts, outperforming state-of-the-art methods. Our code is publicly available at https://github.com/LTS5/ReservoirTTA.

A robust anomaly detection mechanism should possess the capability to effectively remediate anomalies, restoring them to a healthy state, while preserving essential healthy information. Despite the efficacy of existing generative models in learning the underlying distribution of healthy reference data, they face primary challenges when it comes to efficiently repair larger anomalies or anomalies situated near high pixel-density regions. In this paper, we introduce a self-supervised anomaly detection method based on a diffusion model that samples from multi-frequency, four-dimensional simplex noise and makes predictions using our proposed Dynamic Transformer UNet (DTUNet). This simplex-based noise function helps address primary problems to some extent and is scalable for three-dimensional and colored images. In the evolution of ViT, our developed architecture serving as the backbone for the diffusion model, is tailored to treat time and noise image patches as tokens. We incorporate long skip connections bridging the shallow and deep layers, along with smaller skip connections within these layers. Furthermore, we integrate a partial diffusion Markov process, which reduces sampling time, thus enhancing scalability. Our method surpasses existing generative-based anomaly detection methods across three diverse datasets, which include BrainMRI, Brats2021, and the MVtec dataset. It achieves an average improvement of +10.1% in Dice coefficient, +10.4% in IOU, and +9.6% in AUC. Our source code is made publicly available on Github.

In fully supervised learning-based medical image classification, the robustness of a trained model is influenced by its exposure to the range of candidate disease classes. Generalized Zero Shot Learning (GZSL) aims to correctly predict seen and novel unseen classes. Current GZSL approaches have focused mostly on the single-label case. However, it is common for chest X-rays to be labelled with multiple disease classes. We propose a novel multi-modal multi-label GZSL approach that leverages feature disentanglement andmulti-modal information to synthesize features of unseen classes. Disease labels are processed through a pre-trained BioBert model to obtain text embeddings that are used to create a dictionary encoding similarity among different labels. We then use disentangled features and graph aggregation to learn a second dictionary of inter-label similarities. A subsequent clustering step helps to identify representative vectors for each class. The multi-modal multi-label dictionaries and the class representative vectors are used to guide the feature synthesis step, which is the most important component of our pipeline, for generating realistic multi-label disease samples of seen and unseen classes. Our method is benchmarked against multiple competing methods and we outperform all of them based on experiments conducted on the publicly available NIH and CheXpert chest X-ray datasets.

Presents corrections to the paper, (Corrections to "Multi-Label Generalized Zero Shot Chest X-Ray Classification By Combining Image-Text Information With Feature Disentanglement").

Informative sample selection in active learning (AL) helps a machine learning system attain optimum performance with minimum labeled samples, thus improving human-in-the-loop computer-aided diagnosis systems with limited labeled data. Data augmentation is highly effective for enlarging datasets with less labeled data. Combining informative sample selection and data augmentation should leverage their respective advantages and improve performance of AL systems. We propose a novel approach to combine informative sample selection and data augmentation for multi-label active learning. Conventional informative sample selection approaches have mostly focused on the single-label case which do not perform optimally in the multi-label setting. We improve upon state-of-the-art multi-label active learning techniques by representing disease labels as graph nodes, use graph attention transformers (GAT) to learn more effective inter-label relationships and identify most informative samples. We generate transformations of these informative samples which are also informative. Experiments on public chest xray datasets show improved results over state-of-the-art multi-label AL techniques in terms of classification performance, learning rates, and robustness. We also perform qualitative analysis to determine the realism of generated images.

Leveraging nearest neighbor retrieval for self-supervised representation learning has proven beneficial with object-centric images. However, this approach faces limitations when applied to scene-centric datasets, where multiple objects within an image are only implicitly captured in the global representation. Such global bootstrapping can lead to undesirable entanglement of object representations. Furthermore, even object-centric datasets stand to benefit from a finer-grained bootstrapping approach. In response to these challenges, we introduce a novel $\textbf{Cr}$oss-$\textbf{I}$mage Object-Level $\textbf{Bo}$otstrapping method tailored to enhance dense visual representation learning. By employing object-level nearest neighbor bootstrapping throughout the training, CrIBo emerges as a notably strong and adequate candidate for in-context learning, leveraging nearest neighbor retrieval at test time. CrIBo shows state-of-the-art performance on the latter task while being highly competitive in more standard downstream segmentation tasks. Our code and pretrained models are publicly available at https://github.com/tileb1/CrIBo.

Recent test-time adaptation methods heavily rely on nuanced adjustments of batch normalization (BN) parameters. However, one critical assumption often goes overlooked: that of independently and identically distributed (i.i.d.) test batches with respect to unknown labels. This oversight leads to skewed BN statistics and undermines the reliability of the model under non-i.i.d. scenarios. To tackle this challenge, this paper presents a novel method termed '$\textbf{Un-Mix}$ing $\textbf{T}$est-Time $\textbf{N}$ormalization $\textbf{S}$tatistics' (UnMix-TNS). Our method re-calibrates the statistics for each instance within a test batch by $\textit{mixing}$ it with multiple distinct statistics components, thus inherently simulating the i.i.d. scenario. The core of this method hinges on a distinctive online $\textit{unmixing}$ procedure that continuously updates these statistics components by incorporating the most similar instances from new test batches. Remarkably generic in its design, UnMix-TNS seamlessly integrates with a wide range of leading test-time adaptation methods and pre-trained architectures equipped with BN layers. Empirical evaluations corroborate the robustness of UnMix-TNS under varied scenarios—ranging from single to continual and mixed domain shifts, particularly excelling with temporally correlated test data and corrupted non-i.i.d. real-world streams. This adaptability is maintained even with very small batch sizes or single instances. Our results highlight UnMix-TNS's capacity to markedly enhance stability and performance across various benchmarks. Our code is publicly available at https://github.com/devavratTomar/unmixtns.

Developing computational pathology models is essential for reducing manual tissue typing from whole slide images, transferring knowledge from the source domain to an unlabeled, shifted target domain, and identifying unseen categories. We propose a practical setting by addressing the above-mentioned challenges in one fell swoop, i.e., source-free open-set domain adaptation. Our methodology focuses on adapting a pre-trained source model to an unlabeled target dataset and encompasses both closed-set and open-set classes. Beyond addressing the semantic shift of unknown classes, our framework also deals with a covariate shift, which manifests as variations in color appearance between source and target tissue samples. Our method hinges on distilling knowledge from a self-supervised vision transformer (ViT), drawing guidance from either robustly pre-trained transformer models or histopathology datasets, including those from the target domain. In pursuit of this, we introduce a novel style-based adversarial data augmentation, serving as hard positives for self-training a ViT, resulting in highly contextualized embeddings. Following this, we cluster semantically akin target images, with the source model offering weak pseudo-labels, albeit with uncertain confidence. To enhance this process, we present the closed-set affinity score (CSAS), aiming to correct the confidence levels of these pseudo-labels and to calculate weighted class prototypes within the contextualized embedding space. Our approach establishes itself as state-of-the-art across three public histopathological datasets for colorectal cancer assessment. Notably, our self-training method seamlessly integrates with open-set detection methods, resulting in enhanced performance in both closed-set and open-set recognition tasks.

Self-supervised anomaly detection and localization are critical to real-world scenarios in which collecting anomalous samples and pixel-wise labeling is tedious or infeasible, even worse when a wide variety of unseen anomalies could surface at test time. Our approach involves a pretext task in the context of masked image modeling, where the goal is to impose agreement between cluster assignments obtained from the representation of an image view containing saliency-aware masked patches and the uncorrupted image view. We harness the self-attention map extracted from the transformer to mask non-salient image patches without destroying the crucial structure associated with the foreground object. Subsequently, the pre-trained model is fine-tuned to detect and localize simulated anomalies generated under the guidance of the transformer's self-attention map. We conducted extensive validation and ablations on the benchmark of industrial images and achieved superior performance against competing methods. We also show the adaptability of our method to the medical images of the chest X-rays benchmark.

Learning dense visual representations without labels is an arduous task and more so from scene-centric data. We propose to tackle this challenging problem by proposing a Cross-view consistency objective with an Online Clustering mechanism (CrOC) to discover and segment the semantics of the views. In the absence of hand-crafted priors, the resulting method is more generalizable and does not require a cumbersome pre-processing step. More importantly, the clustering algorithm conjointly operates on the features of both views, thereby elegantly bypassing the issue of content not represented in both views and the ambiguous matching of objects from one crop to the other. We demonstrate excellent performance on linear and unsupervised segmentation transfer tasks on various datasets and similarly for video object segmentation. Our code and pre-trained models are publicly available at https://github.com/stegmuel/CrOC.

Most recent test-time adaptation methods focus on only classification tasks, use specialized network architectures, destroy model calibration or rely on lightweight information from the source domain. To tackle these issues, this paper proposes a novel Test-time Self-Learning method with automatic Adversarial augmentation dubbed TeSLA for adapting a pre-trained source model to the unlabeled streaming test data. In contrast to conventional self-learning methods based on cross-entropy, we introduce a new test-time loss function through an implicitly tight connection with the mutual information and online knowledge distillation. Furthermore, we propose a learnable efficient adversarial augmentation module that further enhances online knowledge distillation by simulating high entropy augmented images. Our method achieves state-of-the-art classification and segmentation results on several benchmarks and types of domain shifts, particularly on challenging measurement shifts of medical images. TeSLA also benefits from several desirable properties compared to competing methods in terms of calibration, uncertainty metrics, insensitivity to model architectures, and source training strategies, all supported by extensive ablations. Our code and models are available at https://github.com/devavratTomar/TeSLA.

Most self-supervised methods for representation learning leverage a cross-view consistency objective i.e. they maximize the representation similarity of a given image’s augmented views. Recent work NNCLR goes beyond the cross-view paradigm and uses positive pairs from different images obtained via nearest neighbor bootstrapping in a contrastive setting. We empirically show that as opposed to the contrastive learning setting which relies on negative samples, incorporating nearest neighbor bootstrapping in a self-distillation scheme can lead to a performance drop or even collapse. We scrutinize the reason for this unexpected behavior and provide a solution. We propose to adaptively bootstrap neighbors based on the estimated quality of the latent space. We report consistent improvements compared to the naive bootstrapping approach and the original baselines. Our approach leads to performance improvements for various self-distillation method/backbone combinations and standard downstream tasks. Our code is publicly available at https://github.com/tileb1/AdaSim.

Progress in digital pathology is hindered by high-resolution images and the prohibitive cost of exhaustive localized annotations. The commonly used paradigm to categorize pathology images is patch-based processing, which often incorporates multiple instance learning (MIL) to aggregate local patch-level representations yielding image-level prediction. Nonetheless, diagnostically relevant regions may only take a small fraction of the whole tissue, and current MIL-based approaches often process images uniformly, discarding the inter-patches interactions. To alleviate these issues, we propose ScoreNet, a new efficient transformer that exploits a differentiable recommendation stage to extract discriminative image regions and dedicate computational resources accordingly. The proposed transformer leverages the local and global attention of a few dynamically recommended high-resolution regions at an efficient computational cost. We further introduce a novel mixing data-augmentation, namely ScoreMix, by leveraging the image’s semantic distribution to guide the data mixing and produce coherent sample-label pairs. ScoreMix is embarrassingly simple and mitigates the pitfalls of previous augmentations, which assume a uniform semantic distribution and risk mislabeling the samples. Thorough experiments and ablation studies on three breast cancer histology datasets of Haematoxylin & Eosin (H&E) have validated the superiority of our approach over prior arts, including transformer-based models on tumour regions-of-interest (TRoIs) classification. ScoreNet equipped with proposed ScoreMix augmentation demonstrates better generalization capabilities and achieves new state-of-the-art (SOTA) results with only 50% of the data compared to other mixing augmentation variants. Finally, ScoreNet yields high efficacy and outperforms SOTA efficient transformers, namely TransPath [37] and SwinTransformer [20], with throughput around 3× and 4× higher than the aforementioned architectures, respectively. Our code is publicly available1.

In medical image segmentation, supervised deep networks’ success comes at the cost of requiring abundant labeled data. While asking domain experts to annotate only one or a few of the cohort’s images is feasible, annotating all available images is impractical. This issue is further exacerbated when pre-trained deep networks are exposed to a new image dataset from an unfamiliar distribution. Using available open-source data for ad-hoc transfer learning or hand-tuned techniques for data augmentation only provides suboptimal solutions. Motivated by atlas-based segmentation, we propose a novel volumetric self-supervised learning for data augmentation capable of synthesizing volumetric image-segmentation pairs via learning transformations from a single labeled atlas to the unlabeled data. Our work’s central tenet benefits from a combined view of one-shot generative learning and the proposed self-supervised training strategy that cluster unlabeled volumetric images with similar styles together. Unlike previous methods, our method does not require input volumes at inference time to synthesize new images. Instead, it can generate diversified volumetric image-segmentation pairs from a prior distribution given a single or multi-site dataset. Augmented data generated by our method used to train the segmentation network provide significant improvements over state-of-the-art deep one-shot learning methods on the task of brain MRI segmentation. Ablation studies further exemplified that the proposed appearance model and joint training are crucial to synthesize realistic examples compared to existing medical registration methods. The code, data, and models are available at https://github.com/devavratTomar/SST/.

Supervised learning is constrained by the availability of labeled data, which are especially expensive to acquire in the field of digital pathology. Making use of open-source data for pre-training or using domain adaptation can be a way to overcome this issue. However, pre-trained networks often fail to generalize to new test domains that are not distributed identically due to tissue stainings, types, and textures variations. Additionally, current domain adaptation methods mainly rely on fully-labeled source datasets. In this work, we propose Self-Rule to Multi-Adapt (SRMA), which takes advantage of self-supervised learning to perform domain adaptation, and removes the necessity of fully-labeled source datasets. SRMA can effectively transfer the discriminative knowledge obtained from a few labeled source domain's data to a new target domain without requiring additional tissue annotations. Our method harnesses both domains' structures by capturing visual similarity with intra-domain and cross-domain self-supervision. Moreover, we present a generalized formulation of our approach that allows the framework to learn from multiple source domains. We show that our proposed method outperforms baselines for domain adaptation of colorectal tissue type classification in single and multi-source settings, and further validate our approach on an in-house clinical cohort. The code and trained models are available open-source: https://github.com/christianabbet/SRA.

Explainability of deep learning methods is imperative to facilitate their clinical adoption in digital pathology. However, popular deep learning methods and explainability techniques (explainers) based on pixel-wise processing disregard biological entities’ notion, thus complicating comprehension by pathologists. In this work, we address this by adopting biological entity-based graph processing and graph explainers enabling explanations accessible to pathologists. In this context, a major challenge becomes to discern meaningful explainers, particularly in a standardized and quantifiable fashion. To this end, we propose herein a set of novel quantitative metrics based on statistics of class separability using pathologically measurable concepts to characterize graph explainers. We employ the proposed metrics to evaluate three types of graph explainers, namely the layer-wise relevance propagation, gradient-based saliency, and graph pruning approaches, to explain Cell-Graph representations for Breast Cancer Subtyping. The proposed metrics are also applicable in other domains by using domain-specific intuitive concepts. We validate the qualitative and quantitative findings on the BRACS dataset, a large cohort of breast cancer RoIs, by expert pathologists. The code, data, and models can be accessed here1.

Super-resolution (SR) has traditionally been based on pairs of high-resolution images (HR) and their low-resolution (LR) counterparts obtained artificially with bicubic downsampling. However, in real-world SR, there is a large variety of realistic image degradations and analytically modeling these realistic degradations can prove quite difficult. In this work, we propose to handle real-world SR by splitting this ill-posed problem into two comparatively more well-posed steps. First, we train a network to transform real LR images to the space of bicubically down-sampled images in a supervised manner, by using both real LR/HR pairs and synthetic pairs. Second, we take a generic SR network trained on bicubically downsampled images to super-resolve the transformed LR image. The first step of the pipeline addresses the problem by registering the large variety of degraded images to a common, well understood space of images. The second step then leverages the already impressive performance of SR on bicubically downsampled images, sidestepping the issues of end-to-end training on datasets with many different image degradations. We demonstrate the effectiveness of our proposed method by comparing it to recent methods in real-world SR and show that our proposed approach outperforms the state-of-the-art works in terms of both qualitative and quantitative results, as well as results of an extensive user study conducted on several real image datasets.

Despite the successes of deep neural networks on many challenging vision tasks, they often fail to generalize to new test domains that are not distributed identically to the training data. The domain adaptation becomes more challenging for cross-modality medical data with a notable domain shift. Given that specific annotated imaging modalities may not be accessible nor complete. Our proposed solution is based on the cross-modality synthesis of medical images to reduce the costly annotation burden by radiologists and bridge the domain gap in radiological images. We present a novel approach for image-to-image translation in medical images, capable of supervised or unsupervised (unpaired image data) setups. Built upon adversarial training, we propose a learnable self-attentive spatial normalization of the deep convolutional generator network's intermediate activations. Unlike previous attention-based image-to-image translation approaches, which are either domain-specific or require distortion of the source domain's structures, we unearth the importance of the auxiliary semantic information to handle the geometric changes and preserve anatomical structures during image translation. We achieve superior results for cross-modality segmentation between unpaired MRI and CT data for multi-modality whole heart and multi-modal brain tumor MRI (T1/T2) datasets compared to the state-of-the-art methods. We also observe encouraging results in cross-modality conversion for paired MRI and CT images on a brain dataset. Furthermore, a detailed analysis of the cross-modality image translation, thorough ablation studies confirm our proposed method's efficacy.

Medical image segmentation is important for computer aided diagnosis. Pixelwise manual annotations of large datasets require high expertise and is time consuming. Conventional data augmentations have limited benefit by not fully representing the underlying distribution of the training set, thus affecting model robustness when tested on images captured from different sources. Prior work leverages synthetic images for data augmentation ignoring the interleaved geometric relationship between different anatomical labels. We propose improvements over previous GAN-based medical image synthesis methods by jointly encoding the intrinsic relationship of geometry and shape. Latent space variable sampling results in diverse generated images from a base image and improves robustness. Augmented datasets using our method for automatic segmentation of retinal optical coherence tomography (OCT) images outperform existing methods on the public RETOUCH dataset having images captured from different acquisition procedures. Ablation studies and visual analysis also demonstrate benefits of integrating geometry and diversity.

This paper addresses the problem of automatic facial expression recognition in videos, where the goal is to predict discrete emotion labels best describing the emotions expressed in short video clips. Building on a pre-trained convolutional neural network (CNN) model dedicated to analyzing the video frames and LSTM network designed to process the trajectories of the facial landmarks, this paper investigates several novel directions. First of all, improved face descriptors based on 2D CNNs and facial landmarks are proposed. Second, the paper investigates fusion methods of the features temporally, including a novel hierarchical recurrent neural network combining facial landmark trajectories over time. In addition, we propose a modification to state-of-the-art expression recognition architectures to adapt them to video processing in a simple way. In both ensemble approaches, the temporal information is integrated. Comparative experiments on publicly available video-based facial expression recognition datasets verified that the proposed framework outperforms state-of-the-art methods. Moreover, we introduce a near-infrared video dataset containing facial expressions from subjects driving their cars, which are recorded in real world conditions.

Cross-domain synthesizing realistic faces to learn deep models has attracted increasing attention for facial expression analysis as it helps to improve the performance of expression recognition accuracy despite having small number of real training images. However, learning from synthetic face images can be problematic due to the distribution discrepancy between low-quality synthetic images and real face images and may not achieve the desired performance when the learned model applies to real world scenarios. To this end, we propose a new attribute guided face image synthesis to perform a translation between multiple image domains using a single model. In addition, we adopt the proposed model to learn from synthetic faces by matching the feature distributions between different domains while preserving each domain's characteristics. We evaluate the effectiveness of the proposed approach on several face datasets on generating realistic face images. We demonstrate that the expression recognition performance can be enhanced by benefiting from our face synthesis model. Moreover, we also conduct experiments on a near-infrared dataset containing facial expression videos of drivers to assess the performance using in-the-wild data for driver emotion recognition.

Despite well-established baselines, learning of scene depth and ego-motion from monocular video remains an ongoing challenge, specifically when handling scaling ambiguity issues and depth inconsistencies in image sequences. Much prior work uses either a supervised mode of learning or stereo images. The former is limited by the amount of labeled data, as it requires expensive sensors, while the latter is not always readily available as monocular sequences. In this work, we demonstrate the benefit of using geometric information from synthetic images, coupled with scene depth information, to recover the scale in depth and ego-motion estimation from monocular videos. We developed our framework using synthetic image-depth pairs and unlabeled real monocular images. We had three training objectives: first, to use deep feature alignment to reduce the domain gap between synthetic and monocular images to yield more accurate depth estimation when presented with only real monocular images at test time. Second, we learn scene specific representation by exploiting self-supervision coming from multi-view synthetic images without the need for depth labels. Third, our method uses single-view depth and pose networks, which are capable of jointly training and supervising one another mutually, yielding consistent depth and ego-motion estimates. Extensive experiments demonstrate that our depth and ego-motion models surpass the state-of-the-art, unsupervised methods and compare favorably to early supervised deep models for geometric understanding. We validate the effectiveness of our training objectives against standard benchmarks thorough an ablation study.

By benefiting from perceptual losses, recent studies have improved significantly the performance of the superresolution task, where a high-resolution image is resolved from its low-resolution counterpart. Although such objective functions generate near-photorealistic results, their capability is limited, since they estimate the reconstruction error for an entire image in the same way, without considering any semantic information. In this paper, we propose a novel method to benefit from perceptual loss in a more objective way. We optimize a deep network-based decoder with a targeted objective function that penalizes images at different semantic levels using the corresponding terms. In particular, the proposed method leverages our proposed OBB (Object, Background and Boundary) labels, generated from segmentation labels, to estimate a suitable perceptual loss for boundaries, while considering texture similarity for backgrounds. We show that our proposed approach results in more realistic textures and sharper edges, and outperforms other state-of-the-art algorithms in terms of both qualitative results on standard benchmarks and results of extensive user studies.

Visual tracking of multiple persons simultaneously is an important tool for group behaviour analysis. In this paper, we demonstrate that multi-target tracking in a network of non-overlapping cameras can be formulated in a framework, where the association among all given target hypotheses both within and between cameras is performed simultaneously. Our approach helps to overcome the fragility of multi-camera-based tracking, where the performance relies on the single-camera tracking results obtained at input level. In particular, we formulate an estimation of the target states as a multi-state graph optimization problem, in which the likelihood of each target hypothesis belonging to different identities is modeled. In addition, we learn the target-specific model to improve the similarity measure among targets based on the appearance cues. We also handle the occluded targets when there is no reliable evidence for the target's presence and each target trajectory is expected to be fragmented into multiple tracks. An iterative procedure is proposed to solve the optimization problem, resulting in final trajectories that reveal the true states of the targets. The performance of the proposed approach has been extensively evaluated on challenging multi-camera non-overlapping tracking data sets, in which many difficulties, such as occlusion, viewpoint, and illumination variation, are present. The results of systematic experiments conducted on a large set of sequences show that the proposed approach outperforms several state-of-the-art trackers.

Detecting actions in videos is still a demanding task due to large intra-class variation caused by varying pose, motion and scales. Conventional approaches use a Bag-of-Words model in the form of space-time motion feature pooling followed by learning a classifier. However, since the informative body parts motion only appear in specific regions of the body, these methods have limited capability. In this paper, we seek to learn a model of the interaction among regions of interest via a graph structure. We first discover several space-time video segments representing persistent moving body parts observed sparsely in video. Then, via learning the hidden graph structure (a subset of the graph), we identify both spatial and temporal relations between the subsets of these segments. In order to seize the more discriminative motion patterns and handle different interactions between body parts from simple to composite action, we present a multi-level action model representation. Consequently, for action classification, the classifier learned through each action model labels the test video based on the action model that gives the highest probability score. Experiments on challenging datasets, such as MSR II and UCF-Sports including complex motions and dynamic backgrounds, demonstrate the effectiveness of the proposed approach that outperforms state-of-the-art methods in this context.

In this paper, we propose a new object tracking in a particle filter framework utilising a joint sparsity-based model. Based on the observation that a target can be reconstructed from several templates that are updated dynamically, we jointly analyse the representation of the particles under a single regression framework and with the shared underlying structure. Two convex regularisations are combined and used in our model to enable sparsity as well as facilitate coupling information between particles. Unlike the previous methods that consider a model commonality between particles or regard them as independent tasks, we simultaneously take into account a structure inducing norm and an outlier detecting norm. Such a formulation is shown to be more flexible in terms of handling various types of challenges including occlusion and cluttered background. To derive the optimal solution efficiently, we propose to use a Preconditioned Conjugate Gradient method, which is computationally affordable for high-dimensional data. Furthermore, an online updating procedure scheme is included in the dictionary learning, which makes the proposed tracker less vulnerable to outliers. Experiments on challenging video sequences demonstrate the robustness of the proposed approach to handling occlusion, pose and illumination variation and outperform state-of-the-art trackers in tracking accuracy.