PVTransformer: Point-to-Voxel Transformer for Scalable 3D Object Detection

arXiv:2405.02811v1 Announce Type: new
Abstract: 3D object detectors for point clouds often rely on a pooling-based PointNet to encode sparse points into grid-like voxels or pillars. In this paper, we identify that the common PointNet design introduces an information bottleneck that limits 3D object detection accuracy and scalability. To address this limitation, we propose PVTransformer: a transformer-based point-to-voxel architecture for 3D detection. Our key idea is to replace the PointNet pooling operation with an attention module, leading to a better …

3d object 3d object detection abstract accuracy arxiv cs.cv design detection detectors encode grid identify information object paper pooling scalability scalable transformer type voxel