Learning an egocentric action recognition model from video data is challenging due to distractors (e.g., irrelevant objects) in the background. Further integrating object information into an action model is hence beneficial. Existing methods often leverage a generic object detector to identify and represent the objects in the scene. However, several important issues remain. Object class annotations of good quality for the target domain (dataset) are still required for learning good object representation. Besides, previous methods deeply couple the existing action models and need to retrain them jointly with object representation, leading to costly and inflexible integration. To overcome both limitations, we introduce Self-Supervised Learning Over Sets (SOS), an approach to pre-train a generic Objects In Contact (OIC) representation model from video object regions detected by an off-the-shelf hand-object contact detector. Instead of augmenting object regions individually as in conventional self-supervised learning, we view the action process as a means of natural data transformations with unique spatio-temporal continuity and exploit the inherent relationships among per-video object sets. Extensive experiments on two datasets, EPIC-KITCHENS-100 and EGTEA, show that our OIC significantly boosts the performance of multiple state-of-the-art video classification models.
Temporal action localization (TAL) is an important task extensively explored and improved for third-person videos in recent years. Recent efforts have been made to perform fine-grained temporal localization on first-person videos. However, current TAL methods only use visual signals, neglecting the audio modality that exists in most videos and that shows meaningful action information in egocentric videos. In this work, we take a deep look into the effectiveness of audio in detecting actions in egocentric videos and introduce a simple-yet-effective approach via Observing, Watching, and Listening (OWL) to leverage audio-visual information and context for egocentric TAL. For doing that, we: 1) compare and study different strategies for where and how to fuse the two modalities; 2) propose a transformer-based model to incorporate temporal audio-visual context. Our experiments show that our approach achieves state-of-the-art performance on EPIC-KITCHENS-100.
Continual learning (CL) is under-explored in the video domain. The few existing works contain splits with imbalanced class distributions over the tasks, or study the problem in unsuitable datasets. We introduce vCLIMB, a novel video continual learning benchmark. vCLIMB is a standardized test-bed to analyze catastrophic forgetting of deep models in video continual learning. In contrast to previous work, we focus on class incremental continual learning with models trained on a sequence of disjoint tasks, and distribute the number of classes uniformly across the tasks. We perform in-depth evaluations of existing CL methods in vCLIMB, and observe two unique challenges in video data. The selection of instances to store in episodic memory is performed at the frame level. Second, untrimmed training data influences the effectiveness of frame sampling strategies. We address these two challenges by proposing a temporal consistency regularization that can be applied on top of memory-based continual learning methods. Our approach significantly improves the baseline, by up to 24% on the untrimmed continual learning task. To streamline and foster future research in video continual learning, we will publicly release the code for our benchmark and method.
Recently, few-shot learning has received increasing interest. Existing efforts have been focused on image classification, with very few attempts dedicated to the more challenging few-shot video classification problem. These few attempts aim to effectively exploit the temporal dimension in videos for better learning in low data regimes. However, they have largely ignored a key characteristic of video which could be vital for few-shot recognition, that is, videos are often accompanied by rich text descriptions. In this paper, for the first time, we propose to leverage these human-provided textual descriptions as privileged information when training a few-shot video classification model. Specifically, we formulate a text-based task conditioner to adapt video features to the few-shot learning task. Our model follows a transductive setting where query samples and support textual descriptions can be used to update the support set class prototype to further improve the task-adaptation ability of the model. Our model obtains state-of-the-art performance on four challenging benchmarks in few-shot video action classification.
Many video analysis tasks require temporal localization thus detection of content changes. However, most existing models developed for these tasks are pre-trained on general video action classification tasks. This is because large scale annotation of temporal boundaries in untrimmed videos is expensive. Therefore no suitable datasets exist for temporal boundary-sensitive pre-training. In this paper for the first time, we investigate model pre-training for temporal localization by introducing a novel boundary-sensitive pretext (BSP) task. Instead of relying on costly manual annotations of temporal boundaries, we propose to synthesize temporal boundaries in existing video action classification datasets. With the synthesized boundaries, BSP can be simply conducted via classifying the boundary types. This enables the learning of video representations that are much more transferable to downstream temporal localization tasks. Extensive experiments show that the proposed BSP is superior and complementary to the existing action classification based pre-training counterpart, and achieves new state-of-the-art performance on several temporal localization tasks.
We present the submission of Samsung AI Centre Cambridge to the CVPR2020 EPIC-Kitchens Action Recognition Challenge. In this challenge, action recognition is posed as the problem of simultaneously predicting a single `verb' and `noun' class label given an input trimmed video clip. That is, a `verb' and a `noun' together define a compositional `action' class. The challenging aspects of this real-life action recognition task include small fast moving objects, complex hand-object interactions, and occlusions. At the core of our submission is a recently-proposed spatial-temporal video attention model, called `W3' (`What-Where-When') attention~\cite{perez2020knowing}. We further introduce a simple yet effective contextual learning mechanism to model `action' class scores directly from long-term temporal behaviour based on the `verb' and `noun' prediction scores. Our solution achieves strong performance on the challenge metrics without using object-specific reasoning nor extra training data. In particular, our best solution with multimodal ensemble achieves the 2$^{nd}$ best position for `verb', and 3$^{rd}$ best for `noun' and `action' on the Seen Kitchens test set.
Attentive video modeling is essential for action recognition in unconstrained videos due to their rich yet redundant information over space and time. However, introducing attention in a deep neural network for action recognition is challenging for two reasons. First, an effective attention module needs to learn what (objects and their local motion patterns), where (spatially), and when (temporally) to focus on. Second, a video attention module must be efficient because existing action recognition models already suffer from high computational cost. To address both challenges, a novel What-Where-When (W3) video attention module is proposed. Departing from existing alternatives, our W3 module models all three facets of video attention jointly. Crucially, it is extremely efficient by factorizing the high-dimensional video feature data into low-dimensional meaningful spaces (1D channel vector for `what' and 2D spatial tensors for `where'), followed by lightweight temporal attention reasoning. Extensive experiments show that our attention model brings significant improvements to existing action recognition models, achieving new state-of-the-art performance on a number of benchmarks.
In this paper, we introduce the task of retrieving relevant video moments from a large corpus of untrimmed, unsegmented videos given a natural language query. Our task poses unique challenges as a system must efficiently identify both the relevant videos and localize the relevant moments in the videos. This task is in contrast to prior work that localizes relevant moments in a single video or searches a large collection of already-segmented videos. For our task, we introduce Clip Alignment with Language (CAL), a model that aligns features for a natural language query to a sequence of short video clips that compose a candidate moment in a video. Our approach goes beyond prior work that aggregates video features over a candidate moment by allowing for finer clip alignment. Moreover, our approach is amenable to efficient indexing of the resulting clip-level representations, which makes it suitable for moment localization in large video collections. We evaluate our approach on three recently proposed datasets for temporal localization of moments in video with natural language extended to our video corpus moment retrieval setting: DiDeMo, Charades-STA, and ActivityNet-captions. We show that our CAL model outperforms the recently proposed Moment Context Network (MCN) on all criteria across all datasets on our proposed task, obtaining an 8%-85% and 11%-47% boost for average recall and median rank, respectively, and achieves 5x faster retrieval and 8x smaller index size with a 500K video corpus.