FlexiAvatar: Unified 3D Gaussian Human Avatars Under Arbitrary Body Visibility

2026-07-21

AI Summary

Existing methods for monocular video-driven 3D human avatar generation often introduce artifacts in occluded or invisible body regions due to joint optimization over the entire body surface. This work proposes a unified framework that, for the first time, explicitly optimizes only visible regions by integrating occlusion-robust SMPL-X tracking, part-specific residual geometry and appearance refinement, and diffusion-based texture completion. The approach effectively eliminates artifacts in unobserved areas, achieving an average PSNR improvement of approximately 3% across multiple datasets and significantly outperforming state-of-the-art methods. Furthermore, under partially visible conditions, it reduces the number of Gaussians required, thereby lowering both memory consumption and computational overhead.

Abstract

Reconstructing animatable 3D human avatars from monocular video is a fundamental problem in computer vision with broad applications in AR/VR and digital content creation. Existing approaches typically couple parametric body models with neural rendering or 3D Gaussian splatting and optimize all body regions jointly from short videos, which often degrades fidelity in the visible areas. To overcome this limitation, we introduce FlexiAvatar, a unified framework that explicitly optimizes only the visible body regions, effectively eliminating artifacts arising from unobserved limbs. Our method integrates occlusion-robust SMPL-X tracking with part-specific residual refinement to capture high-frequency geometric and appearance details. To complete entirely unseen regions (e.g., back views), we leverage a diffusion-based approach to generate texture consistent with the observed appearance. Experiments on full-body (NeuMan, ZJU-MoCap, WildAvatar), upper/half-body (talk-show clips), and head-only (INSTA) inputs show that FlexiAvatar delivers consistently higher reconstruction quality, outperforming state-of-the-art methods by an average PSNR improvement of approximately 3% across datasets. Finally, by restricting optimization to observed regions, our method reduces the effective number of Gaussians that must be optimized and rendered, leading to reduced runtime and memory overhead in partial-visibility scenarios.

Problem

• 3D human avatars
• monocular video reconstruction
• body visibility
• occlusion
• fidelity degradation

Innovation

• 3D Gaussian Splatting
• Visible-region Optimization
• Occlusion-robust Tracking
• Diffusion-based Completion
• Animatable Human Avatar