SegVoxelNet: Exploring Semantic Context and Depth-aware Features for 3D Vehicle Detection from Point Cloud | IEEE Conference Publication | IEEE Xplore