Enregistré dans:
Détails bibliographiques
Auteurs principaux: Li, Jiaze, Shi, Yaya, Ma, Zongyang, Xu, Haoran, Cheng, Feng, Xiao, Huihui, Kang, Ruiwen, Yang, Fan, Gao, Tingting, Zhang, Di
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2502.11594
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916618853941248
author Li, Jiaze
Shi, Yaya
Ma, Zongyang
Xu, Haoran
Cheng, Feng
Xiao, Huihui
Kang, Ruiwen
Yang, Fan
Gao, Tingting
Zhang, Di
author_facet Li, Jiaze
Shi, Yaya
Ma, Zongyang
Xu, Haoran
Cheng, Feng
Xiao, Huihui
Kang, Ruiwen
Yang, Fan
Gao, Tingting
Zhang, Di
contents Enhancing the fine-grained instance spatiotemporal motion perception capabilities of Video Large Language Models is crucial for improving their temporal and general video understanding. However, current models struggle to perceive detailed and complex instance motions. To address these challenges, we have made improvements from both data and model perspectives. In terms of data, we have meticulously curated iMOVE-IT, the first large-scale instance-motion-aware video instruction-tuning dataset. This dataset is enriched with comprehensive instance motion annotations and spatiotemporal mutual-supervision tasks, providing extensive training for the model's instance-motion-awareness. Building on this foundation, we introduce iMOVE, an instance-motion-aware video foundation model that utilizes Event-aware Spatiotemporal Efficient Modeling to retain informative instance spatiotemporal motion details while maintaining computational efficiency. It also incorporates Relative Spatiotemporal Position Tokens to ensure awareness of instance spatiotemporal positions. Evaluations indicate that iMOVE excels not only in video temporal understanding and general video understanding but also demonstrates significant advantages in long-term video understanding.
format Preprint
id arxiv_https___arxiv_org_abs_2502_11594
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle iMOVE: Instance-Motion-Aware Video Understanding
Li, Jiaze
Shi, Yaya
Ma, Zongyang
Xu, Haoran
Cheng, Feng
Xiao, Huihui
Kang, Ruiwen
Yang, Fan
Gao, Tingting
Zhang, Di
Computer Vision and Pattern Recognition
Enhancing the fine-grained instance spatiotemporal motion perception capabilities of Video Large Language Models is crucial for improving their temporal and general video understanding. However, current models struggle to perceive detailed and complex instance motions. To address these challenges, we have made improvements from both data and model perspectives. In terms of data, we have meticulously curated iMOVE-IT, the first large-scale instance-motion-aware video instruction-tuning dataset. This dataset is enriched with comprehensive instance motion annotations and spatiotemporal mutual-supervision tasks, providing extensive training for the model's instance-motion-awareness. Building on this foundation, we introduce iMOVE, an instance-motion-aware video foundation model that utilizes Event-aware Spatiotemporal Efficient Modeling to retain informative instance spatiotemporal motion details while maintaining computational efficiency. It also incorporates Relative Spatiotemporal Position Tokens to ensure awareness of instance spatiotemporal positions. Evaluations indicate that iMOVE excels not only in video temporal understanding and general video understanding but also demonstrates significant advantages in long-term video understanding.
title iMOVE: Instance-Motion-Aware Video Understanding
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.11594