Dual-Pathway Sensing with Optimized Attention Network for Video Summarization in Surveillance Systems
ICCK Transactions on Sensing, Communication, and Control · 2025
DOI: 10.62762/tscc.2025.308540 ↗Kurz gefasst
Fachartikel in der Zeitschrift ICCK Transactions on Sensing, Communication, and Control, 2025. Verfasst von Taimur Ali Khan und 3 weiteren Autor:innen.
Abstract (Originalsprache)
Video summarization (VS) aims to generate concise representations of long videos by extracting the most informative frames while maintaining essential content. Existing methods struggle to capture multi-scale dependencies and often rely on suboptimal feature representations, limiting their ability to model complex inter-frame relationships. To address these issues, we propose a multi-scale sensing network that incorporates three key innovations to improve VS. First, we introduce multi-scale dilated convolution blocks with progressively increasing dilation rates to capture temporal context at multiple levels, enabling the network to understand both local transitions and long-range dependencies. Second, we develop a Dual-Pathway Efficient Channel Attention (DECA) module that leverages statistics from Global Average Pooling and Global Max Pooling pathways. Third, we suggest an Optimized Spatial Attention (OSA) module that replaces standard $7\times7$ convolutions with more efficient operations while maintaining spatial dependency modeling. The proposed framework uses EfficientNetB7 as the backbone for robust spatial feature extraction, followed by multi-scale dilated blocks and dual attention mechanisms for detailed feature refinement. Extensive tests on the TVSum and SumMe benchmark datasets demonstrate the superiority of our method, achieving F1 Scores of 63.5% and 53.3%, respectively.
Zitieren
APA
Khan, T. A., Ali, D., Ghazanfar, Z., & Ahmad, B. (2025). Dual-Pathway Sensing with Optimized Attention Network for Video Summarization in Surveillance Systems. ICCK Transactions on Sensing, Communication, and Control. https://doi.org/10.62762/tscc.2025.308540
Vancouver
Khan TA, Ali D, Ghazanfar Z, Ahmad B. Dual-Pathway Sensing with Optimized Attention Network for Video Summarization in Surveillance Systems. ICCK Transactions on Sensing, Communication, and Control. 2025. doi:10.62762/tscc.2025.308540
BibTeX
@article{Khan2025,
author = {Khan, Taimur Ali and Ali, Danish and Ghazanfar, Zainab and Ahmad, Bilal},
title = {Dual-Pathway Sensing with Optimized Attention Network for Video Summarization in Surveillance Systems},
journal = {ICCK Transactions on Sensing, Communication, and Control},
year = {2025},
doi = {10.62762/tscc.2025.308540}
}
EUVIMED