Abstract: We propose DeepPCC, an end-to-end learning-based approach for the lossy compression of large-scale object point clouds. For both geometry and attribute components, we introduce the ...
Abstract: Panoramic segmentation of 3D point clouds is an essential and challenging technology for robots with 3D detection and measurement capabilities. In order to fuse the color information of 2D ...