VS: Reconstructing Clothed 3D Human from Single Image via Vertex Shift

Refine SMPL-X Estimation Strech Figure 1. Using two stages of vertex shift, VS first stretches the estimated SMPL-X into a coarse human model and then refines it into a detailed clothed human model. Results on both benchmarks (1st row) and in-the-wild images (2nd row) show that VS can reconstruct high-fidelity and artifact-less clothed 3D humans from single images, even under scenarios of challenging poses and loose clothes. To comprehensively inspect results from all perspectives, we rotate the results through a full circle and show them from 8 views.