A standard AI vision system assumes every person has four full limbs but a new model finally sees the unique shapes of residual limbs.
Traditional 3D body modeling software relies on a fixed template of the human form that excludes millions of people with limb loss. This system represents the first time a single image can be used to generate a mathematically accurate mesh of a residual limb surface. By adapting the underlying topology of the digital model, it captures the specific contours needed for medical applications. Engineers can now use these 3D reconstructions to design custom prosthetics that fit perfectly without requiring expensive 3D scanners. This technology moves personalized medicine into the realm of a simple smartphone photo.
ResiHMR: Residual-Limb Aware Single-Image 3D Human Mesh Recovery for Individuals with Limb Loss
arXiv · 2604.28025
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 com