Hybrid Causal Action Reasoning for Temporal Action Detection
Open Access
- 27 October 2026
- journal article
- research article
- Published by MDPI AG in Biomimetics
- Vol. 11 (11) , 778
Abstract
Human action understanding aims to detect action instances and classify them in videos. Existing methods capture contextual dependencies through temporal feature aggregation or attention-based interactions. However, these methods primarily rely on visual features to establish relationships, making them susceptible to interference from background content or irrelevant action frames. To address this limitation, we propose Hybrid Causal Action Reasoning (HCAR), a framework for modeling causal correlations between action proposals. Specifically, we design a Hybrid Causality-Aware Attention (HCAA) to refine the semantic representations of proposal classes. To further characterize relations between adjacent proposals, an ordered relation branch evaluates the refined representations in temporal order, while a hyperbolic branch measures hierarchical semantic relatedness between action categories in the Poincaré ball. The outputs of these branches are combined to obtain causal correlation scores, which are subsequently used to refine action proposals. Extensive experiments on THUMOS14 and ActivityNet-1.3 demonstrate the effectiveness of our HCAR.Keywords
Funding Information
- Natural Science Foundation of Shandong Province (ZR2025QC1596)
This publication has 33 references indexed in Scilit:
- CTAA: Cross-Task Attention Alignment for temporal action detectionThe Journal of Supercomputing, 2026
- A novel residual cycle refinement framework for temporal action detection(RCR-TAD)Signal, Image and Video Processing, 2026
- Multi-scale Graph Convolutional Network for understanding human action in videosAdvanced Engineering Informatics, 2025
- DyFADet: Dynamic Feature Aggregation for Temporal Action Detection2024
- Content Temporal Relation Network for temporal action proposal generationPattern Recognition, 2024
- ReAct: Temporal Action Detection with Relational Queries2022
- ActionFormer: Localizing Moments of Actions with Transformers2022
- Predicting memory from the network structure of naturalistic eventsNature Communications, 2022
- BSN: Boundary Sensitive Network for Temporal Action Proposal Generation2018
- Unravelling the contributions of motor experience and conceptual knowledge in action perception: A training studyScientific Reports, 2017