Weakly-supervised temporal action localization (WTAL) learns to detect and classify action instances with only category labels. Most methods widely adopt the off-the-shelf Classification-Based Pre-training (CBP) to generate video features for action localization. However, the different optimization objectives between classification and localization, make temporally localized results suffer from the serious incomplete issue. To tackle this issue without additional annotations, this paper considers to distill free action knowledge from Vision-Language Pre-training (VLP), since we surprisingly observe that the localization results of vanilla VLP have an over-complete issue, which is just complementary to the CBP results. To fuse such complementarity, we propose a novel distillation-collaboration framework with two branches acting as CBP and VLP respectively. The framework is optimized through a dual-branch alternate training strategy. Specifically, during the B step, we distill the confident background pseudo-labels from the CBP branch; while during the F step, the confident foreground pseudo-labels are distilled from the VLP branch. And as a result, the dual-branch complementarity is effectively fused to promote a strong alliance. Extensive experiments and ablation studies on THUMOS14 and ActivityNet1.2 reveal that our method significantly outperforms state-of-the-art methods.
To balance the annotation labor and the granularity of supervision, single-frame annotation has been introduced in temporal action localization. It provides a rough temporal location for an action but implicitly overstates the supervision from the annotated-frame during training, leading to the confusion between actions and backgrounds, i.e., action incompleteness and background false positives. To tackle the two challenges, in this work, we present the Snippet Classification model and the Dilation-Erosion module. In the Dilation-Erosion module, we expand the potential action segments with a loose criterion to alleviate the problem of action incompleteness and then remove the background from the potential action segments to alleviate the problem of action incompleteness. Relying on the single-frame annotation and the output of the snippet classification, the Dilation-Erosion module mines pseudo snippet-level ground-truth, hard backgrounds and evident backgrounds, which in turn further trains the Snippet Classification model. It forms a cyclic dependency. Furthermore, we propose a new embedding loss to aggregate the features of action instances with the same label and separate the features of actions from backgrounds. Experiments on THUMOS14 and ActivityNet 1.2 validate the effectiveness of the proposed method. Code has been made publicly available (https://github.com/LingJun123/single-frame-TAL).
Temporal action localization (TAL) requires long-form reasoning to predict actions of various durations and complex content. Given limited GPU memory, training TAL end to end (i.e., from videos to predictions) on long videos is a significant challenge. Most methods can only train on pre-extracted features without optimizing them for the localization problem, consequently limiting localization performance. In this work, to extend the potential in TAL networks, we propose a novel end-to-end method Re2TAL, which rewires pretrained video backbones for reversible TAL. Re2TAL builds a backbone with reversible modules, where the input can be recovered from the output such that the bulky intermediate activations can be cleared from memory during training. Instead of designing one single type of reversible module, we propose a network rewiring mechanism, to transform any module with a residual connection to a reversible module without changing any parameters. This provides two benefits: (1) a large variety of reversible networks are easily obtained from existing and even future model designs, and (2) the reversible models require much less training effort as they reuse the pre-trained parameters of their original non-reversible versions. Re2TAL, only using the RGB modality, reaches 37.01% average mAP on ActivityNet-v1.3, a new state-of-the-art record, and mAP 64.9% at tIoU=0.5 on THUMOS-14, outperforming all other RGB-only methods.
Weakly supervised temporal action localization (WTAL) aims to localize actions in untrimmed videos with only weak supervision information (e.g. video-level labels). Most existing models handle all input videos with a fixed temporal scale. However, such models are not sensitive to actions whose pace of the movements is different from the ``normal" speed, especially slow-motion action instances, which complete the movements with a much slower speed than their counterparts with a normal speed. Here arises the slow-motion blurred issue: It is hard to explore salient slow-motion information from videos at ``normal" speed. In this paper, we propose a novel framework termed Slow Motion Enhanced Network (SMEN) to improve the ability of a WTAL network by compensating its sensitivity on slow-motion action segments. The proposed SMEN comprises a Mining module and a Localization module. The mining module generates mask to mine slow-motion-related features by utilizing the relationships between the normal motion and slow motion; while the localization module leverages the mined slow-motion features as complementary information to improve the temporal action localization results. Our proposed framework can be easily adapted by existing WTAL networks and enable them be more sensitive to slow-motion actions. Extensive experiments on three benchmarks are conducted, which demonstrate the high performance of our proposed framework.
In this report, we present the ReLER@ZJU1 submission to the Ego4D Moment Queries Challenge in ECCV 2022. In this task, the goal is to retrieve and localize all instances of possible activities in egocentric videos. Ego4D dataset is challenging for the temporal action localization task as the temporal duration of the videos is quite long and each video contains multiple action instances with fine-grained action classes. To address these problems, we utilize a multi-scale transformer to classify different action categories and predict the boundary of each instance. Moreover, in order to better capture the long-term temporal dependencies in the long videos, we propose a segment-level recurrence mechanism. Compared with directly feeding all video features to the transformer encoder, the proposed segment-level recurrence mechanism alleviates the optimization difficulties and achieves better performance. The final submission achieved Recall@1,tIoU=0.5 score of 37.24, average mAP score of 17.67 and took 3-rd place on the leaderboard.
This report describes our submission to the Ego4D Moment Queries Challenge 2022. Our submission builds on ActionFormer, the state-of-the-art backbone for temporal action localization, and a trio of strong video features from SlowFast, Omnivore and EgoVLP. Our solution is ranked 2nd on the public leaderboard with 21.76% average mAP on the test set, which is nearly three times higher than the official baseline. Further, we obtain 42.54% Recall@1x at tIoU=0.5 on the test set, outperforming the top-ranked solution by a significant margin of 1.41 absolute percentage points. Our code is available at https://github.com/happyharrycn/actionformer_release.
This report describes Badgers@UW-Madison, our submission to the Ego4D Natural Language Queries (NLQ) Challenge. Our solution inherits the point-based event representation from our prior work on temporal action localization, and develops a Transformer-based model for video grounding. Further, our solution integrates several strong video features including SlowFast, Omnivore and EgoVLP. Without bells and whistles, our submission based on a single model achieves 12.64% Mean R@1 and is ranked 2nd on the public leaderboard. Meanwhile, our method garners 28.45% (18.03%) R@5 at tIoU=0.3 (0.5), surpassing the top-ranked solution by up to 5.5 absolute percentage points.
Temporal Action Localization (TAL) methods typically operate on top of feature sequences from a frozen snippet encoder that is pretrained with the Trimmed Action Classification (TAC) tasks, resulting in a task discrepancy problem. While existing TAL methods mitigate this issue either by retraining the encoder with a pretext task or by end-to-end fine-tuning, they commonly require an overload of high memory and computation. In this work, we introduce Soft-Landing (SoLa) strategy, an efficient yet effective framework to bridge the transferability gap between the pretrained encoder and the downstream tasks by incorporating a light-weight neural network, i.e., a SoLa module, on top of the frozen encoder. We also propose an unsupervised training scheme for the SoLa module; it learns with inter-frame Similarity Matching that uses the frame interval as its supervisory signal, eliminating the need for temporal annotations. Experimental evaluation on various benchmarks for downstream TAL tasks shows that our method effectively alleviates the task discrepancy problem with remarkable computational efficiency.
Temporal action localization (TAL) aims to detect the boundary and identify the class of each action instance in a long untrimmed video. Current approaches treat video frames homogeneously, and tend to give background and key objects excessive attention. This limits their sensitivity to localize action boundaries. To this end, we propose a prior-enhanced temporal action localization method (PETAL), which only takes in RGB input and incorporates action subjects as priors. This proposal leverages action subjects' information with a plug-and-play subject-aware spatial attention module (SA-SAM) to generate an aggregated and subject-prioritized representation. Experimental results on THUMOS-14 and ActivityNet-1.3 datasets demonstrate that the proposed PETAL achieves competitive performance using only RGB features, e.g., boosting mAP by 2.41% or 0.25% over the state-of-the-art approach that uses RGB features or with additional optical flow features on the THUMOS-14 dataset.
Online Temporal Action Localization (On-TAL) aims to immediately provide action instances from untrimmed streaming videos. The model is not allowed to utilize future frames and any processing techniques to modify past predictions, making On-TAL much more challenging. In this paper, we propose a simple yet effective framework, termed SimOn, that learns to predict action instances using the popular Transformer architecture in an end-to-end manner. Specifically, the model takes the current frame feature as a query and a set of past context information as keys and values of the Transformer. Different from the prior work that uses a set of outputs of the model as past contexts, we leverage the past visual context and the learnable context embedding for the current query. Experimental results on the THUMOS14 and ActivityNet1.3 datasets show that our model remarkably outperforms the previous methods, achieving a new state-of-the-art On-TAL performance. In addition, the evaluation for Online Detection of Action Start (ODAS) demonstrates the effectiveness and robustness of our method in the online setting. The code is available at https://github.com/TuanTNG/SimOn
Understanding human tasks through video observations is an essential capability of intelligent agents. The challenges of such capability lie in the difficulty of generating a detailed understanding of situated actions, their effects on object states (i.e., state changes), and their causal dependencies. These challenges are further aggravated by the natural parallelism from multi-tasking and partial observations in multi-agent collaboration. Most prior works leverage action localization or future prediction as an indirect metric for evaluating such task understanding from videos. To make a direct evaluation, we introduce the EgoTaskQA benchmark that provides a single home for the crucial dimensions of task understanding through question-answering on real-world egocentric videos. We meticulously design questions that target the understanding of (1) action dependencies and effects, (2) intents and goals, and (3) agents' beliefs about others. These questions are divided into four types, including descriptive (what status?), predictive (what will?), explanatory (what caused?), and counterfactual (what if?) to provide diagnostic analyses on spatial, temporal, and causal understandings of goal-oriented tasks. We evaluate state-of-the-art video reasoning models on our benchmark and show their significant gaps between humans in understanding complex goal-oriented egocentric videos. We hope this effort will drive the vision community to move onward with goal-oriented video understanding and reasoning.
Temporal Action Localization (TAL) aims to predict both action category and temporal boundary of action instances in untrimmed videos, i.e., start and end time. Fully-supervised solutions are usually adopted in most existing works, and proven to be effective. One of the practical bottlenecks in these solutions is the large amount of labeled training data required. To reduce expensive human label cost, this paper focuses on a rarely investigated yet practical task named semi-supervised TAL and proposes an effective active learning method, named AL-STAL. We leverage four steps for actively selecting video samples with high informativeness and training the localization model, named \emph{Train, Query, Annotate, Append}. Two scoring functions that consider the uncertainty of localization model are equipped in AL-STAL, thus facilitating the video sample rank and selection. One takes entropy of predicted label distribution as measure of uncertainty, named Temporal Proposal Entropy (TPE). And the other introduces a new metric based on mutual information between adjacent action proposals and evaluates the informativeness of video samples, named Temporal Context Inconsistency (TCI). To validate the effectiveness of proposed method, we conduct extensive experiments on two benchmark datasets THUMOS'14 and ActivityNet 1.3. Experiment results show that AL-STAL outperforms the existing competitors and achieves satisfying performance compared with fully-supervised learning.
Detecting actions in videos have been widely applied in on-device applications. Practical on-device videos are always untrimmed with both action and background. It is desirable for a model to both recognize the class of action and localize the temporal position where the action happens. Such a task is called temporal action location (TAL), which is always trained on the cloud where multiple untrimmed videos are collected and labeled. It is desirable for a TAL model to continuously and locally learn from new data, which can directly improve the action detection precision while protecting customers' privacy. However, it is non-trivial to train a TAL model, since tremendous video samples with temporal annotations are required. However, annotating videos frame by frame is exorbitantly time-consuming and expensive. Although weakly-supervised TAL (W-TAL) has been proposed to learn from untrimmed videos with only video-level labels, such an approach is also not suitable for on-device learning scenarios. In practical on-device learning applications, data are collected in streaming. Dividing such a long video stream into multiple video segments requires lots of human effort, which hinders the exploration of applying the TAL tasks to realistic on-device learning applications. To enable W-TAL models to learn from a long, untrimmed streaming video, we propose an efficient video learning approach that can directly adapt to new environments. We first propose a self-adaptive video dividing approach with a contrast score-based segment merging approach to convert the video stream into multiple segments. Then, we explore different sampling strategies on the TAL tasks to request as few labels as possible. To the best of our knowledge, we are the first attempt to directly learn from the on-device, long video stream.
Temporal action localization aims to predict the boundary and category of each action instance in untrimmed long videos. Most of previous methods based on anchors or proposals neglect the global-local context interaction in entire video sequences. Besides, their multi-stage designs cannot generate action boundaries and categories straightforwardly. To address the above issues, this paper proposes a end-to-end model, called Adaptive Perception transformer (AdaPerFormer for short). Specifically, AdaPerFormer explores a dual-branch attention mechanism. One branch takes care of the global perception attention, which can model entire video sequences and aggregate global relevant contexts. While the other branch concentrates on the local convolutional shift to aggregate intra-frame and inter-frame information through our bidirectional shift operation. The end-to-end nature produces the boundaries and categories of video actions without extra steps. Extensive experiments together with ablation studies are provided to reveal the effectiveness of our design. Our method obtains competitive performance on the THUMOS14 and ActivityNet-1.3 dataset.
Temporal action localization plays an important role in video analysis, which aims to localize and classify actions in untrimmed videos. The previous methods often predict actions on a feature space of a single-temporal scale. However, the temporal features of a low-level scale lack enough semantics for action classification while a high-level scale cannot provide rich details of the action boundaries. To address this issue, we propose to predict actions on a feature space of multi-temporal scales. Specifically, we use refined feature pyramids of different scales to pass semantics from high-level scales to low-level scales. Besides, to establish the long temporal scale of the entire video, we use a spatial-temporal transformer encoder to capture the long-range dependencies of video frames. Then the refined features with long-range dependencies are fed into a classifier for the coarse action prediction. Finally, to further improve the prediction accuracy, we propose to use a frame-level self attention module to refine the classification and boundaries of each action instance. Extensive experiments show that the proposed method can outperform state-of-the-art approaches on the THUMOS14 dataset and achieves comparable performance on the ActivityNet1.3 dataset. Compared with A2Net (TIP20, Avg\{0.3:0.7\}), Sub-Action (CSVT2022, Avg\{0.1:0.5\}), and AFSD (CVPR21, Avg\{0.3:0.7\}) on the THUMOS14 dataset, the proposed method can achieve improvements of 12.6\%, 17.4\% and 2.2\%, respectively
Understanding human emotions is a crucial ability for intelligent robots to provide better human-robot interactions. The existing works are limited to trimmed video-level emotion classification, failing to locate the temporal window corresponding to the emotion. In this paper, we introduce a new task, named Temporal Emotion Localization in videos~(TEL), which aims to detect human emotions and localize their corresponding temporal boundaries in untrimmed videos with aligned subtitles. TEL presents three unique challenges compared to temporal action localization: 1) The emotions have extremely varied temporal dynamics; 2) The emotion cues are embedded in both appearances and complex plots; 3) The fine-grained temporal annotations are complicated and labor-intensive. To address the first two challenges, we propose a novel dilated context integrated network with a coarse-fine two-stream architecture. The coarse stream captures varied temporal dynamics by modeling multi-granularity temporal contexts. The fine stream achieves complex plots understanding by reasoning the dependency between the multi-granularity temporal contexts from the coarse stream and adaptively integrates them into fine-grained video segment features. To address the third challenge, we introduce a cross-modal consensus learning paradigm, which leverages the inherent semantic consensus between the aligned video and subtitle to achieve weakly-supervised learning. We contribute a new testing set with 3,000 manually-annotated temporal boundaries so that future research on the TEL problem can be quantitatively evaluated. Extensive experiments show the effectiveness of our approach on temporal emotion localization. The repository of this work is at https://github.com/YYJMJC/Temporal-Emotion-Localization-in-Videos.
While deep learning has been widely used for video analytics, such as video classification and action detection, dense action detection with fast-moving subjects from sports videos is still challenging. In this work, we release yet another sports video benchmark \TheName{} for \emph{\underline{P}}ing \emph{\underline{P}}ong-\emph{\underline{A}}ction detection, which consists of 2,721 video clips collected from the broadcasting videos of professional table tennis matches in World Table Tennis Championships and Olympiads. We work with a crew of table tennis professionals and referees on a specially designed annotation toolbox to obtain fine-grained action labels (in 14 classes) for every ping-pong action that appeared in the dataset, and formulate two sets of action detection problems -- \emph{action localization} and \emph{action recognition}. We evaluate a number of commonly-seen action recognition (e.g., TSM, TSN, Video SwinTransformer, and Slowfast) and action localization models (e.g., BSN, BSN++, BMN, TCANet), using \TheName{} for both problems, under various settings. These models can only achieve 48\% area under the AR-AN curve for localization and 82\% top-one accuracy for recognition since the ping-pong actions are dense with fast-moving subjects but broadcasting videos are with only 25 FPS. The results confirm that \TheName{} is still a challenging task and can be used as a special benchmark for dense action detection from videos.
Temporal action localization (TAL) is a task of identifying a set of actions in a video, which involves localizing the start and end frames and classifying each action instance. Existing methods have addressed this task by using predefined anchor windows or heuristic bottom-up boundary-matching strategies, which are major bottlenecks in inference time. Additionally, the main challenge is the inability to capture long-range actions due to a lack of global contextual information. In this paper, we present a novel anchor-free framework, referred to as HTNet, which predicts a set of <start time, end time, class> triplets from a video based on a Transformer architecture. After the prediction of coarse boundaries, we refine it through a background feature sampling (BFS) module and hierarchical Transformers, which enables our model to aggregate global contextual information and effectively exploit the inherent semantic relationships in a video. We demonstrate how our method localizes accurate action instances and achieves state-of-the-art performance on two TAL benchmark datasets: THUMOS14 and ActivityNet 1.3.
With video-level labels, weakly supervised temporal action localization (WTAL) applies a localization-by-classification paradigm to detect and classify the action in untrimmed videos. Due to the characteristic of classification, class-specific background snippets are inevitably mis-activated to improve the discriminability of the classifier in WTAL. To alleviate the disturbance of background, existing methods try to enlarge the discrepancy between action and background through modeling background snippets with pseudo-snippet-level annotations, which largely rely on artificial hypotheticals. Distinct from the previous works, we present an adversarial learning strategy to break the limitation of mining pseudo background snippets. Concretely, the background classification loss forces the whole video to be regarded as the background by a background gradient reinforcement strategy, confusing the recognition model. Reversely, the foreground(action) loss guides the model to focus on action snippets under such conditions. As a result, competition between the two classification losses drives the model to boost its ability for action modeling. Simultaneously, a novel temporal enhancement network is designed to facilitate the model to construct temporal relation of affinity snippets based on the proposed strategy, for further improving the performance of action localization. Finally, extensive experiments conducted on THUMOS14 and ActivityNet1.2 demonstrate the effectiveness of the proposed method.
Distracted driving causes thousands of deaths per year, and how to apply deep-learning methods to prevent these tragedies has become a crucial problem. In Track3 of the 6th AI City Challenge, researchers provide a high-quality video dataset with densely action annotations. Due to the small data scale and unclear action boundary, the dataset presents a unique challenge to precisely localize all the different actions and classify their categories. In this paper, we make good use of the multi-view synchronization among videos, and conduct robust Multi-View Practice (MVP) for driving action localization. To avoid overfitting, we fine-tune SlowFast with Kinetics-700 pre-training as the feature extractor. Then the features of different views are passed to ActionFormer to generate candidate action proposals. For precisely localizing all the actions, we design elaborate post-processing, including model voting, threshold filtering and duplication removal. The results show that our MVP is robust for driving action localization, which achieves 28.49% F1-score in the Track3 test set.