Enregistré dans:
Détails bibliographiques
Auteurs principaux: Xi, Ruihao, Wang, Xuekuan, Li, Yongcheng, Li, Shuhua, Wang, Zichen, Wang, Yiwei, Wei, Feng, Zhao, Cairong
Format: Preprint
Publié: 2025
Sujets:
Accès en ligne:https://arxiv.org/abs/2506.14428
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866916796849717248
author Xi, Ruihao
Wang, Xuekuan
Li, Yongcheng
Li, Shuhua
Wang, Zichen
Wang, Yiwei
Wei, Feng
Zhao, Cairong
author_facet Xi, Ruihao
Wang, Xuekuan
Li, Yongcheng
Li, Shuhua
Wang, Zichen
Wang, Yiwei
Wei, Feng
Zhao, Cairong
contents Generating realistic and controllable human motions, particularly those involving rich multi-character interactions, remains a significant challenge due to data scarcity and the complexities of modeling inter-personal dynamics. To address these limitations, we first introduce a new large-scale rich video human motion 2D dataset (Motion2D-Video-150K) comprising 150,000 video sequences. Motion2D-Video-150K features a balanced distribution of diverse single-character and, crucially, double-character interactive actions, each paired with detailed textual descriptions. Building upon this dataset, we propose a novel diffusion-based rich video human motion2D generation (RVHM2D) model. RVHM2D incorporates an enhanced textual conditioning mechanism utilizing either dual text encoders (CLIP-L/B) or T5-XXL with both global and local features. We devise a two-stage training strategy: the model is first trained with a standard diffusion objective, and then fine-tuned using reinforcement learning with an FID-based reward to further enhance motion realism and text alignment. Extensive experiments demonstrate that RVHM2D achieves leading performance on the Motion2D-Video-150K benchmark in generating both single and interactive double-character scenarios.
format Preprint
id arxiv_https___arxiv_org_abs_2506_14428
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Toward Rich Video Human-Motion2D Generation
Xi, Ruihao
Wang, Xuekuan
Li, Yongcheng
Li, Shuhua
Wang, Zichen
Wang, Yiwei
Wei, Feng
Zhao, Cairong
Computer Vision and Pattern Recognition
Generating realistic and controllable human motions, particularly those involving rich multi-character interactions, remains a significant challenge due to data scarcity and the complexities of modeling inter-personal dynamics. To address these limitations, we first introduce a new large-scale rich video human motion 2D dataset (Motion2D-Video-150K) comprising 150,000 video sequences. Motion2D-Video-150K features a balanced distribution of diverse single-character and, crucially, double-character interactive actions, each paired with detailed textual descriptions. Building upon this dataset, we propose a novel diffusion-based rich video human motion2D generation (RVHM2D) model. RVHM2D incorporates an enhanced textual conditioning mechanism utilizing either dual text encoders (CLIP-L/B) or T5-XXL with both global and local features. We devise a two-stage training strategy: the model is first trained with a standard diffusion objective, and then fine-tuned using reinforcement learning with an FID-based reward to further enhance motion realism and text alignment. Extensive experiments demonstrate that RVHM2D achieves leading performance on the Motion2D-Video-150K benchmark in generating both single and interactive double-character scenarios.
title Toward Rich Video Human-Motion2D Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2506.14428