Probabilistic Temporal Masked Attention for Cross-view Online Action Detection

• Published in: IEEE transactions on multimedia pp. 1 - 11
• Main Authors: Xie, Liping, Tan, Yang, Jing, Shicheng, Lu, Huimin, Zhang, Kanjian
• Format: Journal Article
• Language: English
• Published: IEEE 2025
• Subjects: Adaptation models; Autoencoders; Computer architecture; Cross-view online action detection; Encoding; Feature extraction; Microprocessors; Probabilistic logic; Probabilistic modeling; Temporal masked attention; Training; Transformers; Variational autoencoder; Videos
• ISSN: 1520-9210; 1941-0077
• DOI: 10.1109/TMM.2025.3618551

As a critical task in video sequence classification within computer vision, Online Action Detection (OAD) has garnered significant attention. The sensitivity of mainstream OAD models to varying video viewpoints often hampers their generalization when confronted with unseen sources. To address this limitation, we propose a novel Probabilistic Temporal Masked Attention (PTMA) model, which leverages probabilistic modeling to derive latent compressed representations of video frames in a cross-view setting. The PTMA model incorporates a GRU-based temporal masked attention (TMA) cell, which leverages these representations to effectively query the input video sequence, thereby enhancing information interaction and facilitating autoregressive frame-level video analysis. Additionally, multi-view information can be integrated into the probabilistic modeling to facilitate the extraction of view-invariant features. Experiments conducted under three evaluation protocols-cross-subject (cs), cross-view (cv), and cross-subject-view (csv)-demonstrate that the PTMA achieves state-of-the-art performance on the DAHLIA, IKEA ASM, and Breakfast datasets.