Timezone: »

 
Poster
CAGroup3D: Class-Aware Grouping for 3D Object Detection on Point Clouds
Haiyang Wang · Lihe Ding · Shaocong Dong · Shaoshuai Shi · Aoxue Li · Jianan Li · Zhenguo Li · Liwei Wang

@

We present a novel two-stage fully sparse convolutional 3D object detection framework, named CAGroup3D. Our proposed method first generates some high-quality 3D proposals by leveraging the class-aware local group strategy on the object surface voxels with the same semantic predictions, which considers semantic consistency and diverse locality abandoned in previous bottom-up approaches. Then, to recover the features of missed voxels due to incorrect voxel-wise segmentation, we build a fully sparse convolutional RoI pooling module to directly aggregate fine-grained spatial information from backbone for further proposal refinement. It is memory-and-computation efficient and can better encode the geometry-specific features of each 3D proposal. Our model achieves state-of-the-art 3D detection performance with remarkable gains of +3.6% on ScanNet V2 and +2.6% on SUN RGB-D in term of mAP@0.25. Code will be available at https://github.com/Haiyang-W/CAGroup3D.

Keywords: point clouds · 3D object detection · 3D Bounding Boxes · 3D Deep Learning
[ Paper]  [ Poster]  [ OpenReview] 

Author Information

Haiyang Wang (Peking University)
Lihe Ding (BIT)
Shaocong Dong (Beijing Institute of Technology)
Shaoshuai Shi (Saarland Informatics Campus, Max-Planck Institute)
Aoxue Li (Peking University)
Jianan Li (Beijing Institute of Technology)
Zhenguo Li (Noah's Ark Lab, Huawei Tech Investment Co Ltd)
Liwei Wang (Peking University)

More from the Same Authors