Enregistré dans:
Détails bibliographiques
Auteurs principaux: Li, Yuze, Gong, Dong, Cao, Xiao, Yuan, Junchao, Li, Dongsheng, Zhou, Lei, Koh, Yun Sing, Yan, Cheng, Zhang, Xinyu
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2603.01000
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866910051976871936
author Li, Yuze
Gong, Dong
Cao, Xiao
Yuan, Junchao
Li, Dongsheng
Zhou, Lei
Koh, Yun Sing
Yan, Cheng
Zhang, Xinyu
author_facet Li, Yuze
Gong, Dong
Cao, Xiao
Yuan, Junchao
Li, Dongsheng
Zhou, Lei
Koh, Yun Sing
Yan, Cheng
Zhang, Xinyu
contents Motion transfer has emerged as a promising direction for controllable video generation, yet existing methods largely focus on single-object scenarios and struggle when multiple objects require distinct motion patterns. In this work, we present FlexiMMT, the first implicit image-to-video (I2V) motion transfer framework that explicitly enables multi-object, multi-motion transfer. Given a static multi-object image and multiple reference videos, FlexiMMT independently extracts motion representations and accurately assigns them to different objects, supporting flexible recombination and arbitrary motion-to-object mappings. To address the core challenge of cross-object motion entanglement, we introduce a Motion Decoupled Mask Attention Mechanism that uses object-specific masks to constrain attention, ensuring that motion and text tokens only influence their designated regions. We further propose a Differentiated Mask Propagation Mechanism that derives object-specific masks directly from diffusion attention and progressively propagates them across frames efficiently. Extensive experiments demonstrate that FlexiMMT achieves precise, compositional, and state-of-the-art performance in I2V-based multi-object multi-motion transfer. Our project page is: https://ethan-li123.github.io/FlexiMMT_page/
format Preprint
id arxiv_https___arxiv_org_abs_2603_01000
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Let Your Image Move with Your Motion! -- Implicit Multi-Object Multi-Motion Transfer
Li, Yuze
Gong, Dong
Cao, Xiao
Yuan, Junchao
Li, Dongsheng
Zhou, Lei
Koh, Yun Sing
Yan, Cheng
Zhang, Xinyu
Computer Vision and Pattern Recognition
Motion transfer has emerged as a promising direction for controllable video generation, yet existing methods largely focus on single-object scenarios and struggle when multiple objects require distinct motion patterns. In this work, we present FlexiMMT, the first implicit image-to-video (I2V) motion transfer framework that explicitly enables multi-object, multi-motion transfer. Given a static multi-object image and multiple reference videos, FlexiMMT independently extracts motion representations and accurately assigns them to different objects, supporting flexible recombination and arbitrary motion-to-object mappings. To address the core challenge of cross-object motion entanglement, we introduce a Motion Decoupled Mask Attention Mechanism that uses object-specific masks to constrain attention, ensuring that motion and text tokens only influence their designated regions. We further propose a Differentiated Mask Propagation Mechanism that derives object-specific masks directly from diffusion attention and progressively propagates them across frames efficiently. Extensive experiments demonstrate that FlexiMMT achieves precise, compositional, and state-of-the-art performance in I2V-based multi-object multi-motion transfer. Our project page is: https://ethan-li123.github.io/FlexiMMT_page/
title Let Your Image Move with Your Motion! -- Implicit Multi-Object Multi-Motion Transfer
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.01000