Fan Bohao ; Zheng Wenzhao ; Feng Jianjiang ; Zhou Jie

Abstract

In recent years, point cloud perception tasks have been garnering increasingattention. This paper presents the first attempt to estimate 3D human body meshfrom sparse LiDAR point clouds. We found that the major challenge in estimatinghuman pose and mesh from point clouds lies in the sparsity, noise, andincompletion of LiDAR point clouds. Facing these challenges, we propose aneffective sparse-to-dense reconstruction scheme to reconstruct 3D human mesh.This involves estimating a sparse representation of a human (3D human pose) andgradually reconstructing the body mesh. To better leverage the 3D structuralinformation of point clouds, we employ a cascaded graph transformer(graphormer) to introduce point cloud features during sparse-to-densereconstruction. Experimental results on three publicly available databasesdemonstrate the effectiveness of the proposed approach. Code:https://github.com/soullessrobot/LiDAR-HMR/