Srijan Das

Dominick Reilly, Aman Chadha, Srijan Das · apple-ml, stanford

Human perception of surroundings is often guided by the various poses present within the environment. Many computer vision tasks, such as human action recognition and robot imitation learning, rely on pose-based entities like human skeletons or robotic arms. However, conventional Vision Transformer (ViT) models uniformly process all patches, neglecting valuable pose priors in input videos. We argue that incorporating poses into RGB data is advantageous for learning fine-grained and viewpoint-agnostic representations. Consequently, we introduce two strategies for learning pose-aware representations in ViTs. The first method, called Pose-aware Attention Block (PAAB), is a plug-and-play ViT block that performs localized attention on pose regions within videos. The second method, dubbed Pose-Aware Auxiliary Task (PAAT), presents an auxiliary pose prediction task optimized jointly with the primary ViT task. Although their functionalities differ, both methods succeed in learning pose-aware representations, enhancing performance in multiple diverse downstream tasks. Our experiments, conducted across seven datasets, reveal the efficacy of both pose-aware methods on three video analysis tasks, with PAAT holding a slight edge over PAAB. Both PAAT and PAAB surpass their respective backbone Transformers by up to 9.8% in real-world action recognition and 21.8% in multi-view robotic video alignment. Code is available at https://github.com/dominickrei/PoseAwareVT.

Srijan Das, Michael S. Ryoo

In this report, we introduce our adaptation of image-text models for long-term action anticipation. Our Video + CLIP framework makes use of a large-scale pre-trained paired image-text model: CLIP and a video encoder Slowfast network. The CLIP embedding provides fine-grained understanding of objects relevant for an action whereas the slowfast network is responsible for modeling temporal information within a video clip of few frames. We show that the features obtained from both encoders are complementary to each other, thus outperforming the baseline on Ego4D for the task of long-term action anticipation. Our code is available at github.com/srijandas07/clip_baseline_LTA_Ego4d.

Jinghuan Shang, Srijan Das, Michael S. Ryoo

Humans are remarkably flexible in understanding viewpoint changes due to visual cortex supporting the perception of 3D structure. In contrast, most of the computer vision models that learn visual representation from a pool of 2D images often fail to generalize over novel camera viewpoints. Recently, the vision architectures have shifted towards convolution-free architectures, visual Transformers, which operate on tokens derived from image patches. However, these Transformers do not perform explicit operations to learn viewpoint-agnostic representation for visual understanding. To this end, we propose a 3D Token Representation Layer (3DTRL) that estimates the 3D positional information of the visual tokens and leverages it for learning viewpoint-agnostic representations. The key elements of 3DTRL include a pseudo-depth estimator and a learned camera matrix to impose geometric transformations on the tokens, trained in an unsupervised fashion. These enable 3DTRL to recover the 3D positional information of the tokens from 2D patches. In practice, 3DTRL is easily plugged-in into a Transformer. Our experiments demonstrate the effectiveness of 3DTRL in many vision tasks including image classification, multi-view video alignment, and action recognition. The models with 3DTRL outperform their backbone Transformers in all the tasks with minimal added computation. Our code is available at https://github.com/elicassion/3DTRL.

Srijan Das, Tanmay Jain, Dominick Reilly et al.

Vision Transformers (ViTs) have become ubiquitous in computer vision. Despite their success, ViTs lack inductive biases, which can make it difficult to train them with limited data. To address this challenge, prior studies suggest training ViTs with self-supervised learning (SSL) and fine-tuning sequentially. However, we observe that jointly optimizing ViTs for the primary task and a Self-Supervised Auxiliary Task (SSAT) is surprisingly beneficial when the amount of training data is limited. We explore the appropriate SSL tasks that can be optimized alongside the primary task, the training schemes for these tasks, and the data scale at which they can be most effective. Our findings reveal that SSAT is a powerful technique that enables ViTs to leverage the unique characteristics of both the self-supervised and primary tasks, achieving better performance than typical ViTs pre-training with SSL and fine-tuning sequentially. Our experiments, conducted on 10 datasets, demonstrate that SSAT significantly improves ViT performance while reducing carbon footprint. We also confirm the effectiveness of SSAT in the video domain for deepfake detection, showcasing its generalizability. Our code is available at https://github.com/dominickrei/Limited-data-vits.

Xiang Li, Jinghuan Shang, Srijan Das et al.

We investigate whether self-supervised learning (SSL) can improve online reinforcement learning (RL) from pixels. We extend the contrastive reinforcement learning framework (e.g., CURL) that jointly optimizes SSL and RL losses and conduct an extensive amount of experiments with various self-supervised losses. Our observations suggest that the existing SSL framework for RL fails to bring meaningful improvement over the baselines only taking advantage of image augmentation when the same amount of data and augmentation is used. We further perform evolutionary searches to find the optimal combination of multiple self-supervised losses for RL, but find that even such a loss combination fails to meaningfully outperform the methods that only utilize carefully designed image augmentations. After evaluating these approaches together in multiple different environments including a real-world robot environment, we confirm that no single self-supervised loss or image augmentation method can dominate all environments and that the current framework for joint optimization of SSL and RL is limited. Finally, we conduct the ablation study on multiple factors and demonstrate the properties of representations learned with different approaches.

Saarthak Kapse, Srijan Das, Jingwei Zhang et al.

We propose DiRL, a Diversity-inducing Representation Learning technique for histopathology imaging. Self-supervised learning techniques, such as contrastive and non-contrastive approaches, have been shown to learn rich and effective representations of digitized tissue samples with limited pathologist supervision. Our analysis of vanilla SSL-pretrained models' attention distribution reveals an insightful observation: sparsity in attention, i.e, models tends to localize most of their attention to some prominent patterns in the image. Although attention sparsity can be beneficial in natural images due to these prominent patterns being the object of interest itself, this can be sub-optimal in digital pathology; this is because, unlike natural images, digital pathology scans are not object-centric, but rather a complex phenotype of various spatially intermixed biological components. Inadequate diversification of attention in these complex images could result in crucial information loss. To address this, we leverage cell segmentation to densely extract multiple histopathology-specific representations, and then propose a prior-guided dense pretext task for SSL, designed to match the multiple corresponding representations between the views. Through this, the model learns to attend to various components more closely and evenly, thus inducing adequate diversification in attention for capturing context rich representations. Through quantitative and qualitative analysis on multiple tasks across cancer types, we demonstrate the efficacy of our method and observe that the attention is more globally distributed.

Saarthak Kapse, Srijan Das, Prateek Prasanna

Extracting rich phenotype information, such as cell density and arrangement, from whole slide histology images (WSIs), requires analysis of large field of view, i.e more contexual information. This can be achieved through analyzing the digital slides at lower resolution. A potential drawback is missing out on details present at a higher resolution. To jointly leverage complementary information from multiple resolutions, we present a novel transformer based Pyramidal Context-Detail Network (CD-Net). CD-Net exploits the WSI pyramidal structure through co-training of proposed Context and Detail Modules, which operate on inputs from multiple resolutions. The residual connections between the modules enable the joint training paradigm while learning self-supervised representation for WSIs. The efficacy of CD-Net is demonstrated in classifying Lung Adenocarcinoma from Squamous cell carcinoma.

Pranav Balaji, Abhijit Das, Srijan Das et al.

This work explores various ways of exploring multi-task learning (MTL) techniques aimed at classifying videos as original or manipulated in cross-manipulation scenario to attend generalizability in deep fake scenario. The dataset used in our evaluation is FaceForensics++, which features 1000 original videos manipulated by four different techniques, with a total of 5000 videos. We conduct extensive experiments on multi-task learning and contrastive techniques, which are well studied in literature for their generalization benefits. It can be concluded that the proposed detection model is quite generalized, i.e., accurately detects manipulation methods not encountered during training as compared to the state-of-the-art.

Weston Bondurant, Arkaprava Sinha, Hieu Le et al.

Diffusion-based approaches have recently achieved strong results in face swapping, offering improved visual quality over traditional GAN-based methods. However, even state-of-the-art models often suffer from fine-grained artifacts and poor identity preservation, particularly under challenging poses and expressions. A key limitation of existing approaches is their failure to meaningfully leverage 3D facial structure, which is crucial for disentangling identity from pose and expression. In this work, we propose DiffSwap++, a novel diffusion-based face-swapping pipeline that incorporates 3D facial latent features during training. By guiding the generation process with 3D-aware representations, our method enhances geometric consistency and improves the disentanglement of facial identity from appearance attributes. We further design a diffusion architecture that conditions the denoising process on both identity embeddings and facial landmarks, enabling high-fidelity and identity-preserving face swaps. Extensive experiments on CelebA, FFHQ, and CelebV-Text demonstrate that DiffSwap++ outperforms prior methods in preserving source identity while maintaining target pose and expression. Additionally, we introduce a biometric-style evaluation and conduct a user study to further validate the realism and effectiveness of our approach. Code will be made publicly available at https://github.com/WestonBond/DiffSwapPP

Wenhan Wu, Ce Zheng, Zihao Yang et al.

Recently, transformers have demonstrated great potential for modeling long-term dependencies from skeleton sequences and thereby gained ever-increasing attention in skeleton action recognition. However, the existing transformer-based approaches heavily rely on the naive attention mechanism for capturing the spatiotemporal features, which falls short in learning discriminative representations that exhibit similar motion patterns. To address this challenge, we introduce the Frequency-aware Mixed Transformer (FreqMixFormer), specifically designed for recognizing similar skeletal actions with subtle discriminative motions. First, we introduce a frequency-aware attention module to unweave skeleton frequency representations by embedding joint features into frequency attention maps, aiming to distinguish the discriminative movements based on their frequency coefficients. Subsequently, we develop a mixed transformer architecture to incorporate spatial features with frequency features to model the comprehensive frequency-spatial patterns. Additionally, a temporal transformer is proposed to extract the global correlations across frames. Extensive experiments show that FreqMiXFormer outperforms SOTA on 3 popular skeleton action recognition datasets, including NTU RGB+D, NTU RGB+D 120, and NW-UCLA datasets.

Prantik Howlader, Srijan Das, Hieu Le et al.

Incorporating pixel contextual information is critical for accurate segmentation. In this paper, we show that an effective way to incorporate contextual information is through a patch-based classifier. This patch classifier is trained to identify classes present within an image region, which facilitates the elimination of distractors and enhances the classification of small object segments. Specifically, we introduce Multi-scale Patch-based Multi-label Classifier (MPMC), a novel plug-in module designed for existing semi-supervised segmentation (SSS) frameworks. MPMC offers patch-level supervision, enabling the discrimination of pixel regions of different classes within a patch. Furthermore, MPMC learns an adaptive pseudo-label weight, using patch-level classification to alleviate the impact of the teacher's noisy pseudo-label supervision the student. This lightweight module can be integrated into any SSS framework, significantly enhancing their performance. We demonstrate the efficacy of our proposed MPMC by integrating it into four SSS methodologies and improving them across two natural image and one medical segmentation dataset, notably improving the segmentation results of the baselines across all the three datasets.

Dominick Reilly, Srijan Das

Video transformers have become the de facto standard for human action recognition, yet their exclusive reliance on the RGB modality still limits their adoption in certain domains. One such domain is Activities of Daily Living (ADL), where RGB alone is not sufficient to distinguish between visually similar actions, or actions observed from multiple viewpoints. To facilitate the adoption of video transformers for ADL, we hypothesize that the augmentation of RGB with human pose information, known for its sensitivity to fine-grained motion and multiple viewpoints, is essential. Consequently, we introduce the first Pose Induced Video Transformer: PI-ViT (or $π$-ViT), a novel approach that augments the RGB representations learned by video transformers with 2D and 3D pose information. The key elements of $π$-ViT are two plug-in modules, 2D Skeleton Induction Module and 3D Skeleton Induction Module, that are responsible for inducing 2D and 3D pose information into the RGB representations. These modules operate by performing pose-aware auxiliary tasks, a design choice that allows $π$-ViT to discard the modules during inference. Notably, $π$-ViT achieves the state-of-the-art performance on three prominent ADL datasets, encompassing both real-world and large-scale RGB-D datasets, without requiring poses or additional computational overhead at inference.

Varun Belagali, Saarthak Kapse, Pierre Marza et al.

The interpretation of small tiles in large whole slide images (WSI) often needs a larger image context. We introduce TICON, a transformer-based tile representation contextualizer that produces rich, contextualized embeddings for ''any'' application in computational pathology. Standard tile encoder-based pipelines, which extract embeddings of tiles stripped from their context, fail to model the rich slide-level information essential for both local and global tasks. Furthermore, different tile-encoders excel at different downstream tasks. Therefore, a unified model is needed to contextualize embeddings derived from ''any'' tile-level foundation model. TICON addresses this need with a single, shared encoder, pretrained using a masked modeling objective to simultaneously unify and contextualize representations from diverse tile-level pathology foundation models. Our experiments demonstrate that TICON-contextualized embeddings significantly improve performance across many different tasks, establishing new state-of-the-art results on tile-level benchmarks (i.e., HEST-Bench, THUNDER, CATCH) and slide-level benchmarks (i.e., Patho-Bench). Finally, we pretrain an aggregator on TICON to form a slide-level foundation model, using only 11K WSIs, outperforming SoTA slide-level foundation models pretrained with up to 350K WSIs.

Aritra Dutta, Srijan Das, Jacob Nielsen et al.

Despite the commercial abundance of UAVs, aerial data acquisition remains challenging, and the existing Asia and North America-centric open-source UAV datasets are small-scale or low-resolution and lack diversity in scene contextuality. Additionally, the color content of the scenes, solar-zenith angle, and population density of different geographies influence the data diversity. These two factors conjointly render suboptimal aerial-visual perception of the deep neural network (DNN) models trained primarily on the ground-view data, including the open-world foundational models. To pave the way for a transformative era of aerial detection, we present Multiview Aerial Visual RECognition or MAVREC, a video dataset where we record synchronized scenes from different perspectives -- ground camera and drone-mounted camera. MAVREC consists of around 2.5 hours of industry-standard 2.7K resolution video sequences, more than 0.5 million frames, and 1.1 million annotated bounding boxes. This makes MAVREC the largest ground and aerial-view dataset, and the fourth largest among all drone-based datasets across all modalities and tasks. Through our extensive benchmarking on MAVREC, we recognize that augmenting object detectors with ground-view images from the corresponding geographical location is a superior pre-training strategy for aerial detection. Building on this strategy, we benchmark MAVREC with a curriculum-based semi-supervised object detection approach that leverages labeled (ground and aerial) and unlabeled (only aerial) images to enhance the aerial detection. We publicly release the MAVREC dataset: https://mavrec.github.io.

Saarthak Kapse, Pushpak Pati, Srikar Yellapragada et al.

Pretraining a Multiple Instance Learning (MIL) aggregator enables the derivation of Whole Slide Image (WSI)-level embeddings from patch-level representations without supervision. While recent multimodal MIL pretraining approaches leveraging auxiliary modalities have demonstrated performance gains over unimodal WSI pretraining, the acquisition of these additional modalities necessitates extensive clinical profiling. This requirement increases costs and limits scalability in existing WSI datasets lacking such paired modalities. To address this, we propose Gigapixel Vision-Concept Knowledge Contrastive pretraining (GECKO), which aligns WSIs with a Concept Prior derived from the available WSIs. First, we derive an inherently interpretable concept prior by computing the similarity between each WSI patch and textual descriptions of predefined pathology concepts. GECKO then employs a dual-branch MIL network: one branch aggregates patch embeddings into a WSI-level deep embedding, while the other aggregates the concept prior into a corresponding WSI-level concept embedding. Both aggregated embeddings are aligned using a contrastive objective, thereby pretraining the entire dual-branch MIL model. Moreover, when auxiliary modalities such as transcriptomics data are available, GECKO seamlessly integrates them. Across five diverse tasks, GECKO consistently outperforms prior unimodal and multimodal pretraining approaches while also delivering clinically meaningful interpretability that bridges the gap between computational models and pathology expertise. Code is made available at https://github.com/bmi-imaginelab/GECKO

Srijan Das, Rui Dai, Di Yang et al.

Many attempts have been made towards combining RGB and 3D poses for the recognition of Activities of Daily Living (ADL). ADL may look very similar and often necessitate to model fine-grained details to distinguish them. Because the recent 3D ConvNets are too rigid to capture the subtle visual patterns across an action, this research direction is dominated by methods combining RGB and 3D Poses. But the cost of computing 3D poses from RGB stream is high in the absence of appropriate sensors. This limits the usage of aforementioned approaches in real-world applications requiring low latency. Then, how to best take advantage of 3D Poses for recognizing ADL? To this end, we propose an extension of a pose driven attention mechanism: Video-Pose Network (VPN), exploring two distinct directions. One is to transfer the Pose knowledge into RGB through a feature-level distillation and the other towards mimicking pose driven attention through an attention-level distillation. Finally, these two approaches are integrated into a single model, we call VPN++. We show that VPN++ is not only effective but also provides a high speed up and high resilience to noisy Poses. VPN++, with or without 3D Poses, outperforms the representative baselines on 4 public datasets. Code is available at https://github.com/srijandas07/vpnplusplus.

Shukesh Reddy, Nishit Poddar, Srijan Das et al.

In this work, we explore Self-supervised Learning (SSL) as an auxiliary task to blend the texture-based local descriptors into feature modelling for efficient face analysis. Combining a primary task and a self-supervised auxiliary task is beneficial for robust representation. Therefore, we used the SSL task of mask auto-encoder (MAE) as an auxiliary task to reconstruct texture features such as local patterns along with the primary task for robust and unbiased face analysis. We experimented with our hypothesis on three major paradigms of face analysis: face attribute and face-based emotion analysis, and deepfake detection. Our experiment results exhibit that better feature representation can be gleaned from our proposed model for fair and bias-less face analysis.

Saarthak Kapse, Pushpak Pati, Srijan Das et al.

Introducing interpretability and reasoning into Multiple Instance Learning (MIL) methods for Whole Slide Image (WSI) analysis is challenging, given the complexity of gigapixel slides. Traditionally, MIL interpretability is limited to identifying salient regions deemed pertinent for downstream tasks, offering little insight to the end-user (pathologist) regarding the rationale behind these selections. To address this, we propose Self-Interpretable MIL (SI-MIL), a method intrinsically designed for interpretability from the very outset. SI-MIL employs a deep MIL framework to guide an interpretable branch grounded on handcrafted pathological features, facilitating linear predictions. Beyond identifying salient regions, SI-MIL uniquely provides feature-level interpretations rooted in pathological insights for WSIs. Notably, SI-MIL, with its linear prediction constraints, challenges the prevalent myth of an inevitable trade-off between model interpretability and performance, demonstrating competitive results compared to state-of-the-art methods on WSI-level prediction tasks across three cancer types. In addition, we thoroughly benchmark the local and global-interpretability of SI-MIL in terms of statistical analysis, a domain expert study, and desiderata of interpretability, namely, user-friendliness and faithfulness.

Shukesh Reddy, Srijan Das, Abhijit Das

In this work, we attempted to unleash the potential of self-supervised learning as an auxiliary task that can optimise the primary task of generalised deepfake detection. To explore this, we examined different combinations of the training schemes for these tasks that can be most effective. Our findings reveal that fusing the feature representation from self-supervised auxiliary tasks is a powerful feature representation for the problem at hand. Such a representation can leverage the ultimate potential and bring in a unique representation of both the self-supervised and primary tasks, achieving better performance for the primary task. We experimented on a large set of datasets, which includes DF40, FaceForensics++, Celeb-DF, DFD, FaceShifter, UADFV, and our results showed better generalizability on cross-dataset evaluation when compared with current state-of-the-art detectors.

Arkaprava Sinha, Dominick Reilly, Francois Bremond et al.

The introduction of vision-language models like CLIP has enabled the development of foundational video models capable of generalizing to unseen videos and human actions. However, these models are typically trained on web videos, which often fail to capture the challenges present in Activities of Daily Living (ADL) videos. Existing works address ADL-specific challenges, such as similar appearances, subtle motion patterns, and multiple viewpoints, by combining 3D skeletons and RGB videos. However, these approaches are not integrated with language, limiting their ability to generalize to unseen action classes. In this paper, we introduce SKI models, which integrate 3D skeletons into the vision-language embedding space. SKI models leverage a skeleton-language model, SkeletonCLIP, to infuse skeleton information into Vision Language Models (VLMs) and Large Vision Language Models (LVLMs) through collaborative training. Notably, SKI models do not require skeleton data during inference, enhancing their robustness for real-world applications. The effectiveness of SKI models is validated on three popular ADL datasets for zero-shot action recognition and video caption generation tasks.

Dominick Reilly, Manish Kumar Govind, Le Xue et al.

Large Vision-Language Models (VLMs) excel at general visual reasoning tasks but exhibit sharp performance degradation when applied to novel domains with substantial distribution shifts from pretraining data. Existing domain adaptation approaches finetune different VLM components, but this often results in limited domain-specific feature learning or catastrophic forgetting of prior capabilities. To address these issues, we introduce Vision Contextualized Probing (VisCoP), which augments the VLM's vision encoder with a compact set of learnable visual probes. These probes enable efficient domain-specific adaptation with minimal modification to pretrained parameters. We evaluate VisCoP across three challenging domain adaptation settings-cross-view (exocentric to egocentric), cross-modal (RGB to depth), and cross-task (human understanding to robot control). Experiments show that VisCoP consistently outperforms existing adaptation strategies, achieving superior performance on target domains while effectively retaining source-domain knowledge.

Ali K. Rahimian, Manish K. Govind, Subhajit Maity et al.

Vision Transformers and their variants have achieved remarkable success in diverse visual perception tasks. Despite their effectiveness, they suffer from two significant limitations. First, the quadratic computational complexity of multi-head self-attention (MHSA), which restricts scalability to large token counts, and second, a high dependency on large-scale training data to attain competitive performance. In this paper, to address these challenges, we propose a novel sparse self-attention mechanism named Fibottention. Fibottention employs structured sparsity patterns derived from the Wythoff array, enabling an $\mathcal{O}(N \log N)$ computational complexity in self-attention. By design, its sparsity patterns vary across attention heads, which provably reduces redundant pairwise interactions while ensuring sufficient and diverse coverage. This leads to an \emph{inception-like functional diversity} in the attention heads, and promotes more informative and disentangled representations. We integrate Fibottention into standard Transformer architectures and conduct extensive experiments across multiple domains, including image classification, video understanding, and robot learning. Results demonstrate that models equipped with Fibottention either significantly outperform or achieve on-par performance with their dense MHSA counterparts, while leveraging only $2\%$ of all pairwise interactions across self-attention heads in typical settings, $2-6\%$ of the pairwise interactions in self-attention heads, resulting in substantial computational savings. Moreover, when compared to existing sparse attention mechanisms, Fibottention consistently achieves superior results on a FLOP-equivalency basis. Finally, we provide an in-depth analysis of the enhanced feature diversity resulting from our attention design and discuss its implications for efficient representation learning.

Dominick Reilly, Rajatsubhra Chakraborty, Arkaprava Sinha et al.

Current Large Language Vision Models (LLVMs) trained on web videos perform well in general video understanding but struggle with fine-grained details, complex human-object interactions (HOI), and view-invariant representation learning essential for Activities of Daily Living (ADL). This limitation stems from a lack of specialized ADL video instruction-tuning datasets and insufficient modality integration to capture discriminative action representations. To address this, we propose a semi-automated framework for curating ADL datasets, creating ADL-X, a multiview, multimodal RGBS instruction-tuning dataset. Additionally, we introduce LLAVIDAL, an LLVM integrating videos, 3D skeletons, and HOIs to model ADL's complex spatiotemporal relationships. For training LLAVIDAL a simple joint alignment of all modalities yields suboptimal results; thus, we propose a Multimodal Progressive (MMPro) training strategy, incorporating modalities in stages following a curriculum. We also establish ADL MCQ and video description benchmarks to assess LLVM performance in ADL tasks. Trained on ADL-X, LLAVIDAL achieves state-of-the-art performance across ADL benchmarks. Code and data will be made publicly available at: https://adl-x.github.io/.

Rui Dai, Srijan Das, Michael S. Ryoo et al.

The challenge of long-term video understanding remains constrained by the efficient extraction of object semantics and the modelling of their relationships for downstream tasks. Although the CLIP visual features exhibit discriminative properties for various vision tasks, particularly in object encoding, they are suboptimal for long-term video understanding. To address this issue, we present the Attributes-Aware Network (AAN), which consists of two key components: the Attributes Extractor and a Graph Reasoning block. These components facilitate the extraction of object-centric attributes and the modelling of their relationships within the video. By leveraging CLIP features, AAN outperforms state-of-the-art approaches on two popular action detection datasets: Charades and Toyota Smarthome Untrimmed datasets.

Srijan Das, Michael S. Ryoo

Contrastive representation learning of videos highly relies on the availability of millions of unlabelled videos. This is practical for videos available on web but acquiring such large scale of videos for real-world applications is very expensive and laborious. Therefore, in this paper we focus on designing video augmentation for self-supervised learning, we first analyze the best strategy to mix videos to create a new augmented video sample. Then, the question remains, can we make use of the other modalities in videos for data mixing? To this end, we propose Cross-Modal Manifold Cutmix (CMMC) that inserts a video tesseract into another video tesseract in the feature space across two different modalities. We find that our video mixing strategy STC-mix, i.e. preliminary mixing of videos followed by CMMC across different modalities in a video, improves the quality of learned video representations. We conduct thorough experiments for two downstream tasks: action recognition and video retrieval on two small scale video datasets UCF101, and HMDB51. We also demonstrate the effectiveness of our STC-mix on NTU dataset where domain knowledge is limited. We show that the performance of our STC-mix on both the downstream tasks is on par with the other self-supervised approaches while requiring less training data.

Srijan Das, Michael S. Ryoo

Learning self-supervised video representation predominantly focuses on discriminating instances generated from simple data augmentation schemes. However, the learned representation often fails to generalize over unseen camera viewpoints. To this end, we propose ViewCLR, that learns self-supervised video representation invariant to camera viewpoint changes. We introduce a view-generator that can be considered as a learnable augmentation for any self-supervised pre-text tasks, to generate latent viewpoint representation of a video. ViewCLR maximizes the similarities between the latent viewpoint representation with its representation from the original viewpoint, enabling the learned video encoder to generalize over unseen camera viewpoints. Experiments on cross-view benchmark datasets including NTU RGB+D dataset show that ViewCLR stands as a state-of-the-art viewpoint invariant self-supervised method.

Rui Dai, Srijan Das, Kumara Kahatapitiya et al.

Action detection is an essential and challenging task, especially for densely labelled datasets of untrimmed videos. The temporal relation is complex in those datasets, including challenges like composite action, and co-occurring action. For detecting actions in those complex videos, efficiently capturing both short-term and long-term temporal information in the video is critical. To this end, we propose a novel ConvTransformer network for action detection. This network comprises three main components: (1) Temporal Encoder module extensively explores global and local temporal relations at multiple temporal resolutions. (2) Temporal Scale Mixer module effectively fuses the multi-scale features to have a unified feature representation. (3) Classification module is used to learn the instance center-relative position and predict the frame-level classification scores. The extensive experiments on multiple datasets, including Charades, TSU and MultiTHUMOS, confirm the effectiveness of our proposed method. Our network outperforms the state-of-the-art methods on all three datasets.

Rui Dai, Srijan Das, Francois Bremond

Action detection is an essential and challenging task, especially for densely labelled datasets of untrimmed videos. There are many real-world challenges in those datasets, such as composite action, co-occurring action, and high temporal variation of instance duration. For handling these challenges, we propose to explore both the class and temporal relations of detected actions. In this work, we introduce an end-to-end network: Class-Temporal Relational Network (CTRN). It contains three key components: (1) The Representation Transform Module filters the class-specific features from the mixed representations to build graph-structured data. (2) The Class-Temporal Module models the class and temporal relations in a sequential manner. (3) G-classifier leverages the privileged knowledge of the snippet-wise co-occurring action pairs to further improve the co-occurring action detection. We evaluate CTRN on three challenging densely labelled datasets and achieve state-of-the-art performance, reflecting the effectiveness and robustness of our method.

Snehashis Majhi, Srijan Das, Francois Bremond et al.

Anomaly activities such as robbery, explosion, accidents, etc. need immediate actions for preventing loss of human life and property in real world surveillance systems. Although the recent automation in surveillance systems are capable of detecting the anomalies, but they still need human efforts for categorizing the anomalies and taking necessary preventive actions. This is due to the lack of methodology performing both anomaly detection and classification for real world scenarios. Thinking of a fully automatized surveillance system, which is capable of both detecting and classifying the anomalies that need immediate actions, a joint anomaly detection and classification method is a pressing need. The task of joint detection and classification of anomalies becomes challenging due to the unavailability of dense annotated videos pertaining to anomalous classes, which is a crucial factor for training modern deep architecture. Furthermore, doing it through manual human effort seems impossible. Thus, we propose a method that jointly handles the anomaly detection and classification in a single framework by adopting a weakly-supervised learning paradigm. In weakly-supervised learning instead of dense temporal annotations, only video-level labels are sufficient for learning. The proposed model is validated on a large-scale publicly available UCF-Crime dataset, achieving state-of-the-art results.

Rui Dai, Srijan Das, Francois Bremond

In video understanding, most cross-modal knowledge distillation (KD) methods are tailored for classification tasks, focusing on the discriminative representation of the trimmed videos. However, action detection requires not only categorizing actions, but also localizing them in untrimmed videos. Therefore, transferring knowledge pertaining to temporal relations is critical for this task which is missing in the previous cross-modal KD frameworks. To this end, we aim at learning an augmented RGB representation for action detection, taking advantage of additional modalities at training time through KD. We propose a KD framework consisting of two levels of distillation. On one hand, atomic-level distillation encourages the RGB student to learn the sub-representation of the actions from the teacher in a contrastive manner. On the other hand, sequence-level distillation encourages the student to learn the temporal knowledge from the teacher, which consists of transferring the Global Contextual Relations and the Action Boundary Saliency. The result is an Augmented-RGB stream that can achieve competitive performance as the two-stream network while using only RGB at inference time. Extensive experimental analysis shows that our proposed distillation framework is generic and outperforms other popular cross-modal distillation methods in action detection task.

Rui Dai, Srijan Das, Saurav Sharma et al.

Designing activity detection systems that can be successfully deployed in daily-living environments requires datasets that pose the challenges typical of real-world scenarios. In this paper, we introduce a new untrimmed daily-living dataset that features several real-world challenges: Toyota Smarthome Untrimmed (TSU). TSU contains a wide variety of activities performed in a spontaneous manner. The dataset contains dense annotations including elementary, composite activities and activities involving interactions with objects. We provide an analysis of the real-world challenges featured by our dataset, highlighting the open issues for detection algorithms. We show that current state-of-the-art methods fail to achieve satisfactory performance on the TSU dataset. Therefore, we propose a new baseline method for activity detection to tackle the novel challenges provided by our dataset. This method leverages one modality (i.e. optic flow) to generate the attention weights to guide another modality (i.e RGB) to better detect the activity boundaries. This is particularly beneficial to detect activities characterized by high temporal variance. We show that the method we propose outperforms state-of-the-art methods on TSU and on another popular challenging dataset, Charades.

Srijan Das, Saurav Sharma, Rui Dai et al.

In this paper, we focus on the spatio-temporal aspect of recognizing Activities of Daily Living (ADL). ADL have two specific properties (i) subtle spatio-temporal patterns and (ii) similar visual patterns varying with time. Therefore, ADL may look very similar and often necessitate to look at their fine-grained details to distinguish them. Because the recent spatio-temporal 3D ConvNets are too rigid to capture the subtle visual patterns across an action, we propose a novel Video-Pose Network: VPN. The 2 key components of this VPN are a spatial embedding and an attention network. The spatial embedding projects the 3D poses and RGB cues in a common semantic space. This enables the action recognition framework to learn better spatio-temporal features exploiting both modalities. In order to discriminate similar actions, the attention network provides two functionalities - (i) an end-to-end learnable pose backbone exploiting the topology of human body, and (ii) a coupler to provide joint spatio-temporal attention weights across a video. Experiments show that VPN outperforms the state-of-the-art results for action classification on a large scale human activity dataset: NTU-RGB+D 120, its subset NTU-RGB+D 60, a real-world challenging human activity dataset: Toyota Smarthome and a small scale human-object interaction dataset Northwestern UCLA.

Srijan Das, Michal Koperski, Francois Bremond et al.

In this paper, we propose to improve the traditional use of RNNs by employing a many to many model for video classification. We analyze the importance of modeling spatial layout and temporal encoding for daily living action recognition. Many RGB methods focus only on short term temporal information obtained from optical flow. Skeleton based methods on the other hand show that modeling long term skeleton evolution improves action recognition accuracy. In this work, we propose a deep-temporal LSTM architecture which extends standard LSTM and allows better encoding of temporal information. In addition, we propose to fuse 3D skeleton geometry with deep static appearance. We validate our approach on public available CAD60, MSRDailyActivity3D and NTU-RGB+D, achieving competitive performance as compared to the state-of-the art.