Full Professor, Xi'an Jiaotong University
1 paper at NeurIPS 2025
A dual-diffusion pretraining framework for point clouds that predicts patch centers and masked geometry to learn robust semantic and geometric representations.