Abstract
LiDAR point clouds often suffer from sparsity and uneven distributions in outdoor scenes, leading to the poor performance of cutting-edge 3D object detectors. In this paper, we propose Point-Rich, which is designed to improve the performance of 3D object detection. Point-Rich consists of two key modules: HighDensity and HighLight. The HighDensity module addresses the issue of density imbalance by enhancing the point cloud density. The HighLight module leverages image semantic features to enrich the point clouds. Importantly, Point-Rich imposes no restrictions on the 3D object detection architecture and remains unaffected by feature or depth blur. The experimental results show that compared with the Pointpillars on the KITTI dataset, the mAP of Point-Rich under the bird’s eyes view improves by 5.53% on average.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
