ResiHMR: Residual-Limb Aware Single-Image 3D Human Mesh Recovery for Individuals with Limb Loss

arXiv cs.CV / 5/1/2026

📰 NewsIdeas & Deep AnalysisModels & Research

共有:

Key Points

ResiHMR is a new arXiv system for single-image 3D human mesh recovery that specifically targets people with limb loss, overcoming limitations of fixed-topology models that assume intact limbs.
The method uses residual-limb keypoints and adds a topology-adaptive Residual Anchor-Factor Optimization module plus a geometry-based Residual-Limb Reconstruction module to estimate residual-limb boundaries and termination shapes.
By reconstructing residual-limb surfaces and optimizing according to limb-loss topology, ResiHMR aims to better match prosthetic biomechanics and real-world rehabilitation/interaction needs.
Experiments on a real-world curated dataset show improved reconstruction quality across two backbones (SMPLify-X and HSMR), with substantial reductions in both intact-joint and residual-limb 2D MPJPE metrics.
The authors claim this is the first single-image HMR approach that explicitly reconstructs residual-limb surfaces while performing topology-adaptive optimization for limb-loss scenarios.

Abstract

Single-image human mesh recovery provides a compact 3D, person-centric representation that supports analysis, animation, AR and VR, rehabilitation, and human-computer interaction. However, prevailing systems impose an intact-limb prior and degrade on people with limb loss, because fixed-topology models cannot represent residual limbs. In this work, we present ResiHMR, a residual-limb aware framework for single-image 3D human modeling. ResiHMR adopts residual-limb keypoints and introduces two components: (i) a topology-adaptive Residual Anchor-Factor Optimization module that constrains estimation to the observed kinematic subgraph of anatomically valid structures, and (ii) a geometry-based Residual-Limb Reconstruction module that estimates residual-limb boundaries and convex limb-termination geometry. These components introduce topology-aware optimization and explicit termination geometry as tools for human mesh recovery under non-standard limb anatomy. Unlike joint-removal methods in a fixed topology, ResiHMR explicitly reconstructs residual-limb surfaces and aligns optimization with limb-loss topology, which better matches prosthetic biomechanics and real-world use. To the best of our knowledge, this is the first single-image HMR system that explicitly reconstructs residual-limb surfaces and performs topology-adaptive optimization for individuals with limb loss. On a curated dataset of real-world images with limb loss, ResiHMR improves reconstruction quality under both SMPLify-X and HSMR backbones, reducing intact-joint 2D MPJPE from 41.32 to 37.40 with SMPLify-X and residual-limb 2D MPJPE from 73.61 to 23.19 with HSMR.