SAM-Body4D: Training-Free 4D Human Body Mesh Recovery from Videos

TL;DR

SAM-Body4D introduces a training-free method for producing temporally consistent and occlusion-robust 3D human body meshes from videos by leveraging inherent video continuity and a novel occlusion-aware refinement process.

Contribution

It presents a novel training-free framework that enhances temporal stability and occlusion robustness in 3D human mesh recovery from videos without additional training.

Findings

Improved temporal stability in 3D mesh reconstructions

Enhanced robustness to occlusions in challenging videos

Achieves these improvements without retraining the model

Abstract

Human Mesh Recovery (HMR) aims to reconstruct 3D human pose and shape from 2D observations and is fundamental to human-centric understanding in real-world scenarios. While recent image-based HMR methods such as SAM 3D Body achieve strong robustness on in-the-wild images, they rely on per-frame inference when applied to videos, leading to temporal inconsistency and degraded performance under occlusions. We address these issues without extra training by leveraging the inherent human continuity in videos. We propose SAM-Body4D, a training-free framework for temporally consistent and occlusion-robust HMR from videos. We first generate identity-consistent masklets using a promptable video segmentation model, then refine them with an Occlusion-Aware module to recover missing regions. The refined masklets guide SAM 3D Body to produce consistent full-body mesh trajectories, while a padding-based parallel strategy enables efficient multi-human inference. Experimental results demonstrate that SAM-Body4D achieves improved temporal stability and robustness in challenging in-the-wild videos, without any retraining. Our code and demo are available at: https://github.com/gaomingqi/sam-body4d.