Understanding human actions from body poses is critical for assistive robots sharing space with humans in order to make informed and safe decisions about the next interaction. However, precise temporal localization and annotation of activity sequences is time-consuming and the resulting labels are often noisy. If not effectively addressed, label noise negatively affects the model's training, resulting in lower recognition quality. Despite its importance, addressing label noise for skeleton-based action recognition has been overlooked so far. In this study, we bridge this gap by implementing a framework that augments well-established skeleton-based human action recognition methods with label-denoising strategies from various research areas to serve as the initial benchmark. Observations reveal that these baselines yield only marginal performance when dealing with sparse skeleton data. Consequently, we introduce a novel methodology, NoiseEraSAR, which integrates global sample selection, co-teaching, and Cross-Modal Mixture-of-Experts (CM-MOE) strategies, aimed at mitigating the adverse impacts of label noise. Our proposed approach demonstrates better performance on the established benchmark, setting new state-of-the-art standards. The source code for this study is accessible at https://github.com/xuyizdby/NoiseEraSAR.
On modern industrial assembly lines, many intelligent algorithms have been developed to replace or supervise workers. However, we found that there were bottlenecks in both training datasets and real-time performance when deploying algorithms on actual assembly line. Therefore, we developed a promising strategy for expanding industrial datasets, which utilized large models with strong generalization abilities to achieve efficient, high-quality, and large-scale dataset expansion, solving the problem of insufficient and low-quality industrial datasets. We also applied this strategy to video action recognition. We proposed a method of converting hand action recognition problems into hand skeletal trajectory classification problems, which solved the real-time performance problem of industrial algorithms. In the "hand movements during wire insertion" scenarios on the actual assembly line, the accuracy of hand action recognition reached 98.8\%. We conducted detailed experimental analysis to demonstrate the effectiveness and superiority of the method, and deployed the entire process on Midea's actual assembly line.
Understanding videos is one of the fundamental directions in computer vision research, with extensive efforts dedicated to exploring various architectures such as RNN, 3D CNN, and Transformers. The newly proposed architecture of state space model, e.g., Mamba, shows promising traits to extend its success in long sequence modeling to video modeling. To assess whether Mamba can be a viable alternative to Transformers in the video understanding domain, in this work, we conduct a comprehensive set of studies, probing different roles Mamba can play in modeling videos, while investigating diverse tasks where Mamba could exhibit superiority. We categorize Mamba into four roles for modeling videos, deriving a Video Mamba Suite composed of 14 models/modules, and evaluating them on 12 video understanding tasks. Our extensive experiments reveal the strong potential of Mamba on both video-only and video-language tasks while showing promising efficiency-performance trade-offs. We hope this work could provide valuable data points and insights for future research on video understanding. Code is public: https://github.com/OpenGVLab/video-mamba-suite.
Skeleton-based motion representations are robust for action localization and understanding for their invariance to perspective, lighting, and occlusion, compared with images. Yet, they are often ambiguous and incomplete when taken out of context, even for human annotators. As infants discern gestures before associating them with words, actions can be conceptualized before being grounded with labels. Therefore, we propose the first unsupervised pre-training framework, Boundary-Interior Decoding (BID), that partitions a skeleton-based motion sequence into discovered semantically meaningful pre-action segments. By fine-tuning our pre-training network with a small number of annotated data, we show results out-performing SOTA methods by a large margin.
Human action recognition in videos is a critical task with significant implications for numerous applications, including surveillance, sports analytics, and healthcare. The challenge lies in creating models that are both precise in their recognition capabilities and efficient enough for practical use. This study conducts an in-depth analysis of various deep learning models to address this challenge. Utilizing a subset of the UCF101 Videos dataset, we focus on Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Two-Stream ConvNets. The research reveals that while CNNs effectively capture spatial features and RNNs encode temporal sequences, Two-Stream ConvNets exhibit superior performance by integrating spatial and temporal dimensions. These insights are distilled from the evaluation metrics of accuracy, precision, recall, and F1-score. The results of this study underscore the potential of composite models in achieving robust human action recognition and suggest avenues for future research in optimizing these models for real-world deployment.
Temporal localization of driving actions plays a crucial role in advanced driver-assistance systems and naturalistic driving studies. However, this is a challenging task due to strict requirements for robustness, reliability and accurate localization. In this work, we focus on improving the overall performance by efficiently utilizing video action recognition networks and adapting these to the problem of action localization. To this end, we first develop a density-guided label smoothing technique based on label probability distributions to facilitate better learning from boundary video-segments that typically include multiple labels. Second, we design a post-processing step to efficiently fuse information from video-segments and multiple camera views into scene-level predictions, which facilitates elimination of false positives. Our methodology yields a competitive performance on the A2 test set of the naturalistic driving action recognition track of the 2022 NVIDIA AI City Challenge with an F1 score of 0.271.
Classification and localization of driving actions over time is important for advanced driver-assistance systems and naturalistic driving studies. Temporal localization is challenging because it requires robustness, reliability, and accuracy. In this study, we aim to improve the temporal localization and classification accuracy performance by adapting video action recognition and 2D human-pose estimation networks to one model. Therefore, we design a transformer-based fusion architecture to effectively combine 2D-pose features and spatio-temporal features. The model uses 2D-pose features as the positional embedding of the transformer architecture and spatio-temporal features as the main input to the encoder of the transformer. The proposed solution is generic and independent of the camera numbers and positions, giving frame-based class probabilities as output. Finally, the post-processing step combines information from different camera views to obtain final predictions and eliminate false positives. The model performs well on the A2 test set of the 2023 NVIDIA AI City Challenge for naturalistic driving action recognition, achieving the overlap score of the organizer-defined distracted driver behaviour metric of 0.5079.
The increasing variety and quantity of tagged multimedia content on a variety of online platforms offer a unique opportunity to advance the field of human action recognition. In this study, we utilize 283,582 unique, unlabeled TikTok video clips, categorized into 386 hashtags, to train a domain-specific foundation model for action recognition. We employ VideoMAE V2, an advanced model integrating Masked Autoencoders (MAE) with Vision Transformers (ViT), pre-trained on this diverse collection of unstructured videos. Our model, fine-tuned on established action recognition benchmarks such as UCF101 and HMDB51, achieves state-of-the-art results: 99.05% on UCF101, 86.08% on HMDB51, 85.51% on Kinetics-400, and 74.27% on Something-Something V2 using the ViT-giant backbone. These results highlight the potential of using unstructured and unlabeled videos as a valuable source of diverse and dynamic content for training foundation models. Our investigation confirms that while initial increases in pre-training data volume significantly enhance model performance, the gains diminish as the dataset size continues to expand. Our findings emphasize two critical axioms in self-supervised learning for computer vision: (1) additional pre-training data can yield diminishing benefits for some datasets and (2) quality is more important than quantity in self-supervised learning, especially when building foundation models.
Adversarial examples generated by a surrogate model typically exhibit limited transferability to unknown target systems. To address this problem, many transferability enhancement approaches (e.g., input transformation and model augmentation) have been proposed. However, they show poor performances in attacking systems having different model genera from the surrogate model. In this paper, we propose a novel and generic attacking strategy, called Deformation-Constrained Warping Attack (DeCoWA), that can be effectively applied to cross model genus attack. Specifically, DeCoWA firstly augments input examples via an elastic deformation, namely Deformation-Constrained Warping (DeCoW), to obtain rich local details of the augmented input. To avoid severe distortion of global semantics led by random deformation, DeCoW further constrains the strength and direction of the warping transformation by a novel adaptive control strategy. Extensive experiments demonstrate that the transferable examples crafted by our DeCoWA on CNN surrogates can significantly hinder the performance of Transformers (and vice versa) on various tasks, including image classification, video action recognition, and audio recognition. Code is made available at https://github.com/LinQinLiang/DeCoWA.
Data generation is a data augmentation technique for enhancing the generalization ability for skeleton-based human action recognition. Most existing data generation methods face challenges to ensure the temporal consistency of the dynamic information for action. In addition, the data generated by these methods lack diversity when only a few training samples are available. To solve those problems, We propose a novel active generative network (AGN), which can adaptively learn various action categories by motion style transfer to generate new actions when the data for a particular action is only a single sample or few samples. The AGN consists of an action generation network and an uncertainty metric network. The former, with ST-GCN as the Backbone, can implicitly learn the morphological features of the target action while preserving the category features of the source action. The latter guides generating actions. Specifically, an action recognition model generates prediction vectors for each action, which is then scored using an uncertainty metric. Finally, UMN provides the uncertainty sampling basis for the generated actions.
This work explores the performance of a large video understanding foundation model on the downstream task of human fall detection on untrimmed video and leverages a pretrained vision transformer for multi-class action detection, with classes: "Fall", "Lying" and "Other/Activities of daily living (ADL)". A method for temporal action localization that relies on a simple cutup of untrimmed videos is demonstrated. The methodology includes a preprocessing pipeline that converts datasets with timestamp action annotations into labeled datasets of short action clips. Simple and effective clip-sampling strategies are introduced. The effectiveness of the proposed method has been empirically evaluated on the publicly available High-Quality Fall Simulation Dataset (HQFSD). The experimental results validate the performance of the proposed pipeline. The results are promising for real-time application, and the falls are detected on video level with a state-of-the-art 0.96 F1 score on the HQFSD dataset under the given experimental settings. The source code will be made available on GitHub.
Despite recent advances in video action recognition achieving strong performance on existing benchmarks, these models often lack robustness when faced with natural distribution shifts between training and test data. We propose two novel evaluation methods to assess model resilience to such distribution disparity. One method uses two different datasets collected from different sources and uses one for training and validation, and the other for testing. More precisely, we created dataset splits of HMDB-51 or UCF-101 for training, and Kinetics-400 for testing, using the subset of the classes that are overlapping in both train and test datasets. The other proposed method extracts the feature mean of each class from the target evaluation dataset's training data (i.e. class prototype) and estimates test video prediction as a cosine similarity score between each sample to the class prototypes of each target class. This procedure does not alter model weights using the target dataset and it does not require aligning overlapping classes of two different datasets, thus is a very efficient method to test the model robustness to distribution shifts without prior knowledge of the target distribution. We address the robustness problem by adversarial augmentation training - generating augmented views of videos that are "hard" for the classification model by applying gradient ascent on the augmentation parameters - as well as "curriculum" scheduling the strength of the video augmentations. We experimentally demonstrate the superior performance of the proposed adversarial augmentation approach over baselines across three state-of-the-art action recognition models - TSM, Video Swin Transformer, and Uniformer. The presented work provides critical insight into model robustness to distribution shifts and presents effective techniques to enhance video action recognition performance in a real-world deployment.
Recently, the rise of large-scale vision-language pretrained models like CLIP, coupled with the technology of Parameter-Efficient FineTuning (PEFT), has captured substantial attraction in video action recognition. Nevertheless, prevailing approaches tend to prioritize strong supervised performance at the expense of compromising the models' generalization capabilities during transfer. In this paper, we introduce a novel Multimodal, Multi-task CLIP adapting framework named \name to address these challenges, preserving both high supervised performance and robust transferability. Firstly, to enhance the individual modality architectures, we introduce multimodal adapters to both the visual and text branches. Specifically, we design a novel visual TED-Adapter, that performs global Temporal Enhancement and local temporal Difference modeling to improve the temporal representation capabilities of the visual encoder. Moreover, we adopt text encoder adapters to strengthen the learning of semantic label information. Secondly, we design a multi-task decoder with a rich set of supervisory signals to adeptly satisfy the need for strong supervised performance and generalization within a multimodal framework. Experimental results validate the efficacy of our approach, demonstrating exceptional performance in supervised learning while maintaining strong generalization in zero-shot scenarios.
Recent studies have shown that video-level representation learning is crucial to the capture and understanding of the long-range temporal structure for video action recognition. Most existing 3D convolutional neural network (CNN)-based methods for video-level representation learning are clip-based and focus only on short-term motion and appearances. These CNN-based methods lack the capacity to incorporate and model the long-range spatiotemporal representation of the underlying video and ignore the long-range video-level context during training. In this study, we propose a factorized 4D CNN architecture with attention (F4D) that is capable of learning more effective, finer-grained, long-term spatiotemporal video representations. We demonstrate that the proposed F4D architecture yields significant performance improvements over the conventional 2D, and 3D CNN architectures proposed in the literature. Experiment evaluation on five action recognition benchmark datasets, i.e., Something-Something-v1, SomethingSomething-v2, Kinetics-400, UCF101, and HMDB51 demonstrate the effectiveness of the proposed F4D network architecture for video-level action recognition.
This paper proposes a framework that combines online human state estimation, action recognition and motion prediction to enable early assessment and prevention of worker biomechanical risk during lifting tasks. The framework leverages the NIOSH index to perform online risk assessment, thus fitting real-time applications. In particular, the human state is retrieved via inverse kinematics/dynamics algorithms from wearable sensor data. Human action recognition and motion prediction are achieved by implementing an LSTM-based Guided Mixture of Experts architecture, which is trained offline and inferred online. With the recognized actions, a single lifting activity is divided into a series of continuous movements and the Revised NIOSH Lifting Equation can be applied for risk assessment. Moreover, the predicted motions enable anticipation of future risks. A haptic actuator, embedded in the wearable system, can alert the subject of potential risk, acting as an active prevention device. The performance of the proposed framework is validated by executing real lifting tasks, while the subject is equipped with the iFeel wearable system.
Re-identifying participants in ultra-distance running competitions can be daunting due to the extensive distances and constantly changing terrain. To overcome these challenges, computer vision techniques have been developed to analyze runners' faces, numbers on their bibs, and clothing. However, our study presents a novel gait-based approach for runners' re-identification (re-ID) by leveraging various pre-trained human action recognition (HAR) models and loss functions. Our results show that this approach provides promising results for re-identifying runners in ultra-distance competitions. Furthermore, we investigate the significance of distinct human body movements when athletes are approaching their endurance limits and their potential impact on re-ID accuracy. Our study examines how the recognition of a runner's gait is affected by a competition's critical point (CP), defined as a moment of severe fatigue and the point where the finish line comes into view, just a few kilometers away from this location. We aim to determine how this CP can improve the accuracy of athlete re-ID. Our experimental results demonstrate that gait recognition can be significantly enhanced (up to a 9% increase in mAP) as athletes approach this point. This highlights the potential of utilizing gait recognition in real-world scenarios, such as ultra-distance competitions or long-duration surveillance tasks.
This research addresses whether the ball's direction after a soccer free-kick can be accurately predicted solely by observing the shooter's kicking technique. To investigate this, we meticulously curated a dataset of soccer players executing free kicks and conducted manual temporal segmentation to identify the moment of the kick precisely. Our approach involves utilizing neural networks to develop a model that integrates Human Action Recognition (HAR) embeddings with contextual information, predicting the ball-on-goal position (BoGP) based on two temporal states: the kicker's run-up and the instant of the kick. The study encompasses a performance evaluation for eleven distinct HAR backbones, shedding light on their effectiveness in BoGP estimation during free-kick situations. An extra tabular metadata input is introduced, leading to an interesting model enhancement without introducing bias. The promising results reveal 69.1% accuracy when considering two primary BoGP classes: right and left. This underscores the model's proficiency in predicting the ball's destination towards the goal with high accuracy, offering promising implications for understanding free-kick dynamics in soccer.
Skeleton-based action recognition is a central task in human-computer interaction. However, most previous methods suffer from two issues: (i) semantic ambiguity arising from spatial-temporal information mixture; and (ii) overlooking the explicit exploitation of the latent data distributions (i.e., the intra-class variations and inter-class relations), thereby leading to sub-optimum solutions of the skeleton encoders. To mitigate this, we propose a spatial-temporal decoupling contrastive learning (STD-CL) framework to obtain discriminative and semantically distinct representations from the sequences, which can be incorporated into various previous skeleton encoders and can be removed when testing. Specifically, we decouple the global features into spatial-specific and temporal-specific features to reduce the spatial-temporal coupling of features. Furthermore, to explicitly exploit the latent data distributions, we employ the attentive features to contrastive learning, which models the cross-sequence semantic relations by pulling together the features from the positive pairs and pushing away the negative pairs. Extensive experiments show that STD-CL with four various skeleton encoders (HCN, 2S-AGCN, CTR-GCN, and Hyperformer) achieves solid improvements on NTU60, NTU120, and NW-UCLA benchmarks. The code will be released soon.
Temporal Action Localization (TAL) is a complex task that poses relevant challenges, particularly when attempting to generalize on new -- unseen -- domains in real-world applications. These scenarios, despite realistic, are often neglected in the literature, exposing these solutions to important performance degradation. In this work, we tackle this issue by introducing, for the first time, an approach for Unsupervised Domain Adaptation (UDA) in sparse TAL, which we refer to as Semantic Adversarial unsupervised Domain Adaptation (SADA). Our contributions are threefold: (1) we pioneer the development of a domain adaptation model that operates on realistic sparse action detection benchmarks; (2) we tackle the limitations of global-distribution alignment techniques by introducing a novel adversarial loss that is sensitive to local class distributions, ensuring finer-grained adaptation; and (3) we present a novel set of benchmarks based on EpicKitchens100 and CharadesEgo, that evaluate multiple domain shifts in a comprehensive manner. Our experiments indicate that SADA improves the adaptation across domains when compared to fully supervised state-of-the-art and alternative UDA methods, attaining a performance boost of up to 6.14% mAP.
Weakly-supervised temporal action localization aims to localize action instances in videos with only video-level action labels. Existing methods mainly embrace a localization-by-classification pipeline that optimizes the snippet-level prediction with a video classification loss. However, this formulation suffers from the discrepancy between classification and detection, resulting in inaccurate separation of foreground and background (F\&B) snippets. To alleviate this problem, we propose to explore the underlying structure among the snippets by resorting to unsupervised snippet clustering, rather than heavily relying on the video classification loss. Specifically, we propose a novel clustering-based F\&B separation algorithm. It comprises two core components: a snippet clustering component that groups the snippets into multiple latent clusters and a cluster classification component that further classifies the cluster as foreground or background. As there are no ground-truth labels to train these two components, we introduce a unified self-labeling mechanism based on optimal transport to produce high-quality pseudo-labels that match several plausible prior distributions. This ensures that the cluster assignments of the snippets can be accurately associated with their F\&B labels, thereby boosting the F\&B separation. We evaluate our method on three benchmarks: THUMOS14, ActivityNet v1.2 and v1.3. Our method achieves promising performance on all three benchmarks while being significantly more lightweight than previous methods. Code is available at https://github.com/Qinying-Liu/CASE