PV-RCNN_Point-Voxel-Feature-Set-Abstraction-For-3D-Object-Detection
PointVoxel-RCNN (PV-RCNN), is a two-stage 3D detection framework aiming at more accurate 3D object detection from point clouds. 3D detection approaches are based on either 3D voxel CNN with sparse convolution or PointNet-based networks as the backbone. 3D voxel CNNs with sparse convolution are more efficient and are able to generate high-quality 3D object proposals. PointNet-based methods can capture more accurate contextual information with flexible receptive fields.