PointRCNN: 3D Object Proposal Generation and Detection from Point Cloud

CVPR 2019  ·  Shaoshuai Shi, Xiaogang Wang, Hongsheng Li ·

In this paper, we propose PointRCNN for 3D object detection from raw point cloud. The whole framework is composed of two stages: stage-1 for the bottom-up 3D proposal generation and stage-2 for refining proposals in the canonical coordinates to obtain the final detection results. Instead of generating proposals from RGB image or projecting point cloud to bird's view or voxels as previous methods do, our stage-1 sub-network directly generates a small number of high-quality 3D proposals from point cloud in a bottom-up manner via segmenting the point cloud of the whole scene into foreground points and background. The stage-2 sub-network transforms the pooled points of each proposal to canonical coordinates to learn better local spatial features, which is combined with global semantic features of each point learned in stage-1 for accurate box refinement and confidence prediction. Extensive experiments on the 3D detection benchmark of KITTI dataset show that our proposed architecture outperforms state-of-the-art methods with remarkable margins by using only point cloud as input. The code is available at https://github.com/sshaoshuai/PointRCNN.

PDF Abstract CVPR 2019 PDF CVPR 2019 Abstract

Datasets


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
3D Object Detection KITTI Cars Easy PointRCNN AP 84.32% # 17
Object Detection KITTI Cars Easy PointRCNN Shi et al. (2019) AP 85.94 # 2
Object Detection KITTI Cars Hard PointRCNN Shi et al. (2019) AP 68.32 # 2
3D Object Detection KITTI Cars Hard PointRCNN AP 67.86% # 17
3D Object Detection KITTI Cars Moderate PointRCNN AP 75.42% # 20
Object Detection KITTI Cars Moderate PointRCNN Shi et al. (2019) AP 75.76 # 2
3D Object Detection KITTI Cyclists Easy PointRCNN AP 73.93% # 8
3D Object Detection KITTI Cyclists Hard PointRCNN AP 53.59% # 7
3D Object Detection KITTI Cyclists Moderate PointRCNN AP 59.60% # 7

Methods


No methods listed for this paper. Add relevant methods here