Guardado en:
Detalles Bibliográficos
Autores principales: Wang, Bin, Li, Wentong, Wang, Wenqian, Gao, Mingliang, Cong, Runmin, Zhang, Wei
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2408.10688
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908405335064576
author Wang, Bin
Li, Wentong
Wang, Wenqian
Gao, Mingliang
Cong, Runmin
Zhang, Wei
author_facet Wang, Bin
Li, Wentong
Wang, Wenqian
Gao, Mingliang
Cong, Runmin
Zhang, Wei
contents Recently, large-scale pre-trained vision-language models (e.g., CLIP), have garnered significant attention thanks to their powerful representative capabilities. This inspires researchers in transferring the knowledge from these large pre-trained models to other task-specific models, e.g., Video Action Recognition (VAR) models, via particularly leveraging side networks to enhance the efficiency of parameter-efficient fine-tuning (PEFT). However, current transferring approaches in VAR tend to directly transfer the frozen knowledge from large pre-trained models to action recognition networks with minimal cost, instead of exploiting the temporal modeling capabilities of the action recognition models themselves. Therefore, in this paper, we propose a novel memory-efficient Temporal Difference Side Network (TDS-CLIP) to balance knowledge transferring and temporal modeling, avoiding backpropagation in frozen parameter models. Specifically, we introduce a Temporal Difference Adapter (TD-Adapter), which can effectively capture local temporal differences in motion features to strengthen the model's global temporal modeling capabilities. Furthermore, we designed a Side Motion Enhancement Adapter (SME-Adapter) to guide the proposed side network in efficiently learning the rich motion information in videos, thereby improving the side network's ability to capture and learn motion information. Extensive experiments are conducted on three benchmark datasets, including Something-Something V1&V2, and Kinetics-400. Experimental results show that our method achieves competitive performance in video action recognition tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2408_10688
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle TDS-CLIP: Temporal Difference Side Network for Efficient VideoAction Recognition
Wang, Bin
Li, Wentong
Wang, Wenqian
Gao, Mingliang
Cong, Runmin
Zhang, Wei
Computer Vision and Pattern Recognition
Recently, large-scale pre-trained vision-language models (e.g., CLIP), have garnered significant attention thanks to their powerful representative capabilities. This inspires researchers in transferring the knowledge from these large pre-trained models to other task-specific models, e.g., Video Action Recognition (VAR) models, via particularly leveraging side networks to enhance the efficiency of parameter-efficient fine-tuning (PEFT). However, current transferring approaches in VAR tend to directly transfer the frozen knowledge from large pre-trained models to action recognition networks with minimal cost, instead of exploiting the temporal modeling capabilities of the action recognition models themselves. Therefore, in this paper, we propose a novel memory-efficient Temporal Difference Side Network (TDS-CLIP) to balance knowledge transferring and temporal modeling, avoiding backpropagation in frozen parameter models. Specifically, we introduce a Temporal Difference Adapter (TD-Adapter), which can effectively capture local temporal differences in motion features to strengthen the model's global temporal modeling capabilities. Furthermore, we designed a Side Motion Enhancement Adapter (SME-Adapter) to guide the proposed side network in efficiently learning the rich motion information in videos, thereby improving the side network's ability to capture and learn motion information. Extensive experiments are conducted on three benchmark datasets, including Something-Something V1&V2, and Kinetics-400. Experimental results show that our method achieves competitive performance in video action recognition tasks.
title TDS-CLIP: Temporal Difference Side Network for Efficient VideoAction Recognition
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2408.10688