Abstract: The conversion of raw point clouds into pillar representations has been widely adopted for 3D object detection. Such conversion allows a point cloud to be discretized into structured grids, ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results