Research on Temporal Action Localization Methods for Cross-modal Understanding

Authors

• Jinwei Li, College of Systems Engineering, National University of Defense Technology, Changsha 410073, China
• Xiaoyang Liu, College of Systems Engineering, National University of Defense Technology, Changsha 410073, China
• Rusheng Ju, College of Systems Engineering, National University of Defense Technology, Changsha 410073, China

Abstract

Abstract: To address the problems of insufficient localization accuracy and high model complexity in the temporal action localization (TAL) task for video-text cross-modal understanding, an anchor-free action transformer (AFAT) model was proposed. Based on the anchor-free framework, the local self-attention mechanism of Transformer was introduced to enhance the global modeling capability of temporal features. A multi-scale feature pyramid structure was combined to strengthen the representation of actions with different durations, and a lightweight predictor was adopted to reduce computational redundancy. Experimental results show that the average precision of this model on the THUMOS14 dataset is significantly improved compared with the baseline model, and the model complexity is effectively reduced while maintaining high detection performance.

Recommended Citation

Li, Jinwei; Liu, Xiaoyang; and Ju, Rusheng (2026) "Research on Temporal Action Localization Methods for Cross-modal Understanding," Journal of System Simulation: Vol. 38: Iss. 7, Article 18.
DOI: 10.16182/j.issn1004731x.joss.25-0879
Available at: https://dc-china-simulation.researchcommons.org/journal/vol38/iss7/18

First Page

2053

Last Page

2067

CLC

TP391.9

Recommended Citation

Li Jinwei, Liu Xiaoyang, Ju Rusheng. Research on Temporal Action Localization Methods for Cross- modal Understanding[J]. Journal of System Simulation, 2026, 38(7): 2053-2067.

Corresponding Author

Ju Rusheng

DOI

10.16182/j.issn1004731x.joss.25-0879