ImVideoEdit: Image-learning Video Editing via 2D Spatial Difference Attention Blocks

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Xu, Jiayang, Zhuo, Fan, Zhang, Majun, Pan, Changhao, Wang, Zehan, Chen, Siyu, Yang, Xiaoda, Jin, Tao, Zhao, Zhou
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866911618155151360
author Xu, Jiayang
Zhuo, Fan
Zhang, Majun
Pan, Changhao
Wang, Zehan
Chen, Siyu
Yang, Xiaoda
Jin, Tao
Zhao, Zhou
author_facet Xu, Jiayang
Zhuo, Fan
Zhang, Majun
Pan, Changhao
Wang, Zehan
Chen, Siyu
Yang, Xiaoda
Jin, Tao
Zhao, Zhou
contents Current video editing models often rely on expensive paired video data, which limits their practical scalability. In essence, most video editing tasks can be formulated as a decoupled spatiotemporal process, where the temporal dynamics of the pretrained model are preserved while spatial content is selectively and precisely modified. Based on this insight, we propose ImVideoEdit, an efficient framework that learns video editing capabilities entirely from image pairs. By freezing the pre-trained 3D attention modules and treating images as single-frame videos, we decouple the 2D spatial learning process to help preserve the original temporal dynamics. The core of our approach is a Predict-Update Spatial Difference Attention module that progressively extracts and injects spatial differences. Rather than relying on rigid external masks, we incorporate a Text-Guided Dynamic Semantic Gating mechanism for adaptive and implicit text-driven modifications. Despite training on only 13K image pairs for 5 epochs with exceptionally low computational overhead, ImVideoEdit achieves editing fidelity and temporal consistency comparable to larger models trained on extensive video datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2604_07958
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ImVideoEdit: Image-learning Video Editing via 2D Spatial Difference Attention Blocks
Xu, Jiayang
Zhuo, Fan
Zhang, Majun
Pan, Changhao
Wang, Zehan
Chen, Siyu
Yang, Xiaoda
Jin, Tao
Zhao, Zhou
Computer Vision and Pattern Recognition
Current video editing models often rely on expensive paired video data, which limits their practical scalability. In essence, most video editing tasks can be formulated as a decoupled spatiotemporal process, where the temporal dynamics of the pretrained model are preserved while spatial content is selectively and precisely modified. Based on this insight, we propose ImVideoEdit, an efficient framework that learns video editing capabilities entirely from image pairs. By freezing the pre-trained 3D attention modules and treating images as single-frame videos, we decouple the 2D spatial learning process to help preserve the original temporal dynamics. The core of our approach is a Predict-Update Spatial Difference Attention module that progressively extracts and injects spatial differences. Rather than relying on rigid external masks, we incorporate a Text-Guided Dynamic Semantic Gating mechanism for adaptive and implicit text-driven modifications. Despite training on only 13K image pairs for 5 epochs with exceptionally low computational overhead, ImVideoEdit achieves editing fidelity and temporal consistency comparable to larger models trained on extensive video datasets.
title ImVideoEdit: Image-learning Video Editing via 2D Spatial Difference Attention Blocks
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.07958