How to leverage the temporal dimension is a key question in video analysis. Recent works suggest an efficient approach to video feature learning, i.e., factorizing 3D convolutions into separate components respectively for spatial and temporal convolutions. The temporal convolution, however, comes with an implicit assumption - the feature maps across time steps are well aligned so that the features at the same locations can be aggregated. This assumption may be overly strong in practical applications, especially in action recognition where the motion serves as a crucial cue. In this work, we propose a new CNN architecture TrajectoryNet, which incorporates trajectory convolution, a new operation for integrating features along the temporal dimension, to replace the existing temporal convolution. This operation explicitly takes into account the changes in contents caused by deformation or motion, allowing the visual features to be aggregated along the the motion paths, trajectories. On two large-scale action recognition datasets, namely, Something-Something and Kinetics, the proposed network architecture achieves notable improvement over strong baselines.

State-space models have gained popularity in sequence modelling due to their simple and efficient network structures. However, the absence of nonlinear activation along the temporal direction limits the model's capacity.

How to leverage the temporal dimension is a key question in video analysis. Recent works suggest an efficient approach to video feature learning, i.e., factorizing 3D convolutions into separate components respectively for spatial and temporal convolutions. The temporal convolution, however, comes with an implicit assumption - the feature maps across time steps are well aligned so that the features at the same locations can be aggregated. This assumption may be overly strong in practical applications, especially in action recognition where the motion serves as a crucial cue. In this work, we propose a new CNN architecture TrajectoryNet, which incorporates trajectory convolution, a new operation for integrating features along the temporal dimension, to replace the existing temporal convolution. This operation explicitly takes into account the changes in contents caused by deformation or motion, allowing the visual features to be aggregated along the the motion paths, trajectories. On two large-scale action recognition datasets, namely, Something-Something and Kinetics, the proposed network architecture achieves notable improvement over strong baselines.

This paper presents a framework that improves two stream networks for video action recognition by extending residual network to combine information from two streams into one single network. It significantly improves over previous state-of-the-art on two popular video action recognition benchmark. The downside of this paper is the limited novelty. There are previous work tried to combine two streams into a single network [1,2], and the temporal convolution is not new either [3]. Although the way to combine two streams is slightly different from previous work, the proposed approach is still pretty straightforward.

Skeleton-based action recognition has achieved remarkable performance with the development of graph convolutional networks (GCNs). However, most of these methods tend to construct complex topology learning mechanisms while neglecting the inherent symmetry of the human body. Additionally, the use of temporal convolutions with certain fixed receptive fields limits their capacity to effectively capture dependencies in time sequences. To address the issues, we (1) propose a novel Topological Symmetry Enhanced Graph Convolution (TSE-GC) to enable distinct topology learning across different channel partitions while incorporating topological symmetry awareness and (2) construct a Multi-Branch Deformable Temporal Convolution (MBDTC) for skeleton-based action recognition. The proposed TSE-GC emphasizes the inherent symmetry of the human body while enabling efficient learning of dynamic topologies. Meanwhile, the design of MBDTC introduces the concept of deformable modeling, leading to more flexible receptive fields and stronger modeling capacity of temporal dependencies. Combining TSE-GC with MBDTC, our final model, TSE-GCN, achieves competitive performance with fewer parameters compared with state-of-the-art methods on three large datasets, NTU RGB+D, NTU RGB+D 120, and NW-UCLA. On the cross-subject and cross-set evaluations of NTU RGB+D 120, the accuracies of our model reach 90.0\% and 91.1\%, with 1.1M parameters and 1.38 GFLOPS for one stream.

Ultrasound based hand movement estimation is a crucial area of research with applications in human-machine interaction. Forearm ultrasound offers detailed information about muscle morphology changes during hand movement which can be used to estimate hand gestures. Previous work has focused on analyzing 2-Dimensional (2D) ultrasound image frames using techniques such as convolutional neural networks (CNNs). However, such 2D techniques do not capture temporal features from segments of ultrasound data corresponding to continuous hand movements. This study uses 3D CNN based techniques to capture spatio-temporal patterns within ultrasound video segments for gesture recognition. We compared the performance of a 2D convolution-based network with (2+1)D convolution-based, 3D convolution-based, and our proposed network. Our methodology enhanced the gesture classification accuracy to 98.8 +/- 0.9%, from 96.5 +/- 2.3% compared to a network trained with 2D convolution layers. These results demonstrate the advantages of using ultrasound video snippets for improving hand gesture classification performance.

Event-based data are commonly encountered in edge computing environments where efficiency and low latency are critical. To interface with such data and leverage their rich temporal features, we propose a causal spatiotemporal convolutional network. This solution targets efficient implementation on edge-appropriate hardware with limited resources in three ways: 1) deliberately targets a simple architecture and set of operations (convolutions, ReLU activations) 2) can be configured to perform online inference efficiently via buffering of layer outputs 3) can achieve more than 90% activation sparsity through regularization during training, enabling very significant efficiency gains on event-based processors. In addition, we propose a general affine augmentation strategy acting directly on the events, which alleviates the problem of dataset scarcity for event-based systems. We apply our model on the AIS 2024 event-based eye tracking challenge, reaching a score of 0.9916 p10 accuracy on the Kaggle private testset.

Anomaly detection in dynamic graphs presents a significant challenge due to the temporal evolution of graph structures and attributes. The conventional approaches that tackle this problem typically employ an unsupervised learning framework, capturing normality patterns with exclusive normal data during training and identifying deviations as anomalies during testing. However, these methods face critical drawbacks: they either only depend on proxy tasks for general representation without directly pinpointing normal patterns, or they neglect to differentiate between spatial and temporal normality patterns, leading to diminished efficacy in anomaly detection. To address these challenges, we introduce a novel Spatial-Temporal memories-enhanced graph autoencoder (STRIPE). Initially, STRIPE employs Graph Neural Networks (GNNs) and gated temporal convolution layers to extract spatial features and temporal features, respectively. Then STRIPE incorporates separate spatial and temporal memory networks, which capture and store prototypes of normal patterns, thereby preserving the uniqueness of spatial and temporal normality. After that, through a mutual attention mechanism, these stored patterns are then retrieved and integrated with encoded graph embeddings. Finally, the integrated features are fed into the decoder to reconstruct the graph streams which serve as the proxy task for anomaly detection. This comprehensive approach not only minimizes reconstruction errors but also refines the model by emphasizing the compactness and distinctiveness of the embeddings in relation to the nearest memory prototypes. Through extensive testing, STRIPE has demonstrated a superior capability to discern anomalies by effectively leveraging the distinct spatial and temporal dynamics of dynamic graphs, significantly outperforming existing methodologies, with an average improvement of 15.39% on AUC values.