Revisiting Point Sampling in LiDAR-based 3D Object Detection
Xiaolong Zhao, Wei Tian · 2022 6th CAA International Conference on Vehicular Control and Intelligence (CVCI) · 2022
LiDAR-based 3D object detection has been rapidly developed in many fields, such as robotics and autonomous driving. Mainstream LiDAR-based 3D object detection includes various sampling methods that attempt to reduce computational costs by reaching sufficient accuracy via fewer points. To stimulate related researches, this paper reviews sampling methods in LiDAR-based 3D object detection. It covers both traditional sampling methods and deep learning sampling methods.