PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection
International Journal of Computer Vision · 2023-02 · 131 (2) · 531-551
DOI: 10.1007/s11263-022-01710-9 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Computer Vision, online erschienen am 24.11.2022. Verfasst von Shaoshuai Shi und 7 weiteren Autor:innen. Die Arbeit wurde bisher 494-mal zitiert (Crossref) und nennt 82 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Max Planck Institute for Informatics.
Abstract (Originalsprache)
3D object detection is receiving increasing attention from both industry and academia thanks to its wide applications in various fields. In this paper, we propose Point-Voxel Region-based Convolution Neural Networks (PV-RCNNs) for 3D object detection on point clouds. First, we propose a novel 3D detector, PV-RCNN, which boosts the 3D detection performance by deeply integrating the feature learning of both point-based set abstraction and voxel-based sparse convolution through two novel steps, i.e. , the voxel-to-keypoint scene encoding and the keypoint-to-grid RoI feature abstraction. Second, we propose an advanced framework, PV-RCNN++, for more efficient and accurate 3D object detection. It consists of two major improvements: sectorized proposal-centric sampling for efficiently producing more representative keypoints, and VectorPool aggregation for better aggregating local point features with much less resource consumption. With these two strategies, our PV-RCNN++ is about $$3\times $$ 3 × faster than PV-RCNN, while also achieving better performance. The experiments demonstrate that our proposed PV-RCNN++ framework achieves state-of-the-art 3D detection performance on the large-scale and highly-competitive Waymo Open Dataset with 10 FPS inference speed on the detection range of $$150m \times 150m$$ 150 m × 150 m .
Förderung
- Max Planck Institute for Informatics
Zitieren
APA
Shi, S., Jiang, L., Deng, J., Wang, Z., Guo, C., Shi, J., Wang, X., & Li, H. (2023). PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection. International Journal of Computer Vision, 131(2), 531-551. https://doi.org/10.1007/s11263-022-01710-9
Vancouver
Shi S, Jiang L, Deng J, Wang Z, Guo C, Shi J, et al. PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection. International Journal of Computer Vision. 2023;131(2):531-551. doi:10.1007/s11263-022-01710-9
BibTeX
@article{Shi2023,
author = {Shi, Shaoshuai and Jiang, Li and Deng, Jiajun and Wang, Zhe and Guo, Chaoxu and Shi, Jianping and Wang, Xiaogang and Li, Hongsheng},
title = {PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection},
journal = {International Journal of Computer Vision},
year = {2023},
volume = {131},
number = {2},
pages = {531--551},
doi = {10.1007/s11263-022-01710-9}
}
EUVIMED