Unsupervised Domain Adaptation on Point Clouds via High-Order Geometric Structure Modeling

Information Science, Computer Science, Artificial Intelligence,
point clouds   unsupervised domain adaptation   high-order geometric structures   contrastive learning   self-supervised learning   3D vision   feature alignment  

High-Order Geometric Structure Modeling-Based Unsupervised Domain Adaptation for Point Clouds Research Background and Motivation Point cloud data is a key data form for describing three-dimensional spaces, widely used in real-world applications such as autonomous driving and remote sensing. Point clouds can capture precise geometric information, bu...