Fine-Grained Spatiotemporal Motion Alignment for Contrastive Video Representation Learning

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zhu, Minghao, Lin, Xiao, Dang, Ronghao, Liu, Chengju, Chen, Qijun
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912072804073472
author Zhu, Minghao
Lin, Xiao
Dang, Ronghao
Liu, Chengju
Chen, Qijun
author_facet Zhu, Minghao
Lin, Xiao
Dang, Ronghao
Liu, Chengju
Chen, Qijun
contents As the most essential property in a video, motion information is critical to a robust and generalized video representation. To inject motion dynamics, recent works have adopted frame difference as the source of motion information in video contrastive learning, considering the trade-off between quality and cost. However, existing works align motion features at the instance level, which suffers from spatial and temporal weak alignment across modalities. In this paper, we present a \textbf{Fi}ne-grained \textbf{M}otion \textbf{A}lignment (FIMA) framework, capable of introducing well-aligned and significant motion information. Specifically, we first develop a dense contrastive learning framework in the spatiotemporal domain to generate pixel-level motion supervision. Then, we design a motion decoder and a foreground sampling strategy to eliminate the weak alignments in terms of time and space. Moreover, a frame-level motion contrastive loss is presented to improve the temporal diversity of the motion features. Extensive experiments demonstrate that the representations learned by FIMA possess great motion-awareness capabilities and achieve state-of-the-art or competitive results on downstream tasks across UCF101, HMDB51, and Diving48 datasets. Code is available at \url{https://github.com/ZMHH-H/FIMA}.
format Preprint
id arxiv_https___arxiv_org_abs_2309_00297
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Fine-Grained Spatiotemporal Motion Alignment for Contrastive Video Representation Learning
Zhu, Minghao
Lin, Xiao
Dang, Ronghao
Liu, Chengju
Chen, Qijun
Computer Vision and Pattern Recognition
As the most essential property in a video, motion information is critical to a robust and generalized video representation. To inject motion dynamics, recent works have adopted frame difference as the source of motion information in video contrastive learning, considering the trade-off between quality and cost. However, existing works align motion features at the instance level, which suffers from spatial and temporal weak alignment across modalities. In this paper, we present a \textbf{Fi}ne-grained \textbf{M}otion \textbf{A}lignment (FIMA) framework, capable of introducing well-aligned and significant motion information. Specifically, we first develop a dense contrastive learning framework in the spatiotemporal domain to generate pixel-level motion supervision. Then, we design a motion decoder and a foreground sampling strategy to eliminate the weak alignments in terms of time and space. Moreover, a frame-level motion contrastive loss is presented to improve the temporal diversity of the motion features. Extensive experiments demonstrate that the representations learned by FIMA possess great motion-awareness capabilities and achieve state-of-the-art or competitive results on downstream tasks across UCF101, HMDB51, and Diving48 datasets. Code is available at \url{https://github.com/ZMHH-H/FIMA}.
title Fine-Grained Spatiotemporal Motion Alignment for Contrastive Video Representation Learning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2309.00297