{"ID":2850207,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.22140","arxiv_id":"2510.22140","title":"STG-Avatar: Animatable Human Avatars via Spacetime Gaussian","abstract":"Realistic animatable human avatars from monocular videos are crucial for advancing human-robot interaction and enhancing immersive virtual experiences. While recent research on 3DGS-based human avatars has made progress, it still struggles with accurately representing detailed features of non-rigid objects (e.g., clothing deformations) and dynamic regions (e.g., rapidly moving limbs). To address these challenges, we present STG-Avatar, a 3DGS-based framework for high-fidelity animatable human avatar reconstruction. Specifically, our framework introduces a rigid-nonrigid coupled deformation framework that synergistically integrates Spacetime Gaussians (STG) with linear blend skinning (LBS). In this hybrid design, LBS enables real-time skeletal control by driving global pose transformations, while STG complements it through spacetime adaptive optimization of 3D Gaussians. Furthermore, we employ optical flow to identify high-dynamic regions and guide the adaptive densification of 3D Gaussians in these regions. Experimental results demonstrate that our method consistently outperforms state-of-the-art baselines in both reconstruction quality and operational efficiency, achieving superior quantitative metrics while retaining real-time rendering capabilities. Our code is available at https://github.com/jiangguangan/STG-Avatar","short_abstract":"Realistic animatable human avatars from monocular videos are crucial for advancing human-robot interaction and enhancing immersive virtual experiences. While recent research on 3DGS-based human avatars has made progress, it still struggles with accurately representing detailed features of non-rigid objects (e.g., cloth...","url_abs":"https://arxiv.org/abs/2510.22140","url_pdf":"https://arxiv.org/pdf/2510.22140v1","authors":"[\"Guangan Jiang\",\"Tianzi Zhang\",\"Dong Li\",\"Zhenjun Zhao\",\"Haoang Li\",\"Mingrui Li\",\"Hongyu Wang\"]","published":"2025-10-25T03:23:38Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Generative Adversarial Network\"]","has_code":false,"code_links":[{"ID":607776,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2850207,"paper_url":"https://arxiv.org/abs/2510.22140","paper_title":"STG-Avatar: Animatable Human Avatars via Spacetime Gaussian","repo_url":"https://github.com/jiangguangan/STG-Avatar","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}