Details
| Original language | English |
|---|---|
| Title of host publication | 2023 26th International Conference on Information Fusion, FUSION 2023 |
| ISBN (electronic) | 979-8-89034-485-4 |
| Publication status | Published - 2023 |
| Event | 26th International Conference on Information Fusion, FUSION 2023 - Charleston, United States Duration: 27 Jun 2023 → 30 Jun 2023 |
Abstract
Architectures that first convert point clouds to a grid representation and then apply convolutional neural networks achieve good performance for radar-based object detection. However, the transfer from irregular point cloud data to a dense grid structure is often associated with a loss of information, due to the discretization and aggregation of points. In this paper, we propose a novel architecture, multi-scale KPPillarsBEV, that aims to mitigate the negative effects of grid rendering. Specifically, we propose a novel grid rendering method, KPBEV, which leverages the descriptive power of kernel point convolutions to improve the encoding of local point cloud contexts during grid rendering. In addition, we propose a general multi-scale grid rendering formulation to incorporate multi-scale feature maps into convolutional backbones of detection networks with arbitrary grid rendering methods. We perform extensive experiments on the nuScenes dataset and evaluate the methods in terms of detection performance and computational complexity. The proposed multi-scale KPPillarsBEV architecture outperforms the baseline by 5.37% and the previous state of the art by 2.88% in Car AP4.0 (average precision for a matching threshold of 4 meters) on the nuScenes validation set. Moreover, the proposed single-scale KPBEV grid rendering improves the Car AP4.0 by 2.90% over the baseline while maintaining the same inference speed.
Keywords
- cs.CV, cs.AI, cs.LG, cs.RO
ASJC Scopus subject areas
- Computer Science(all)
- Computer Networks and Communications
- Computer Science(all)
- Computer Vision and Pattern Recognition
- Computer Science(all)
- Signal Processing
- Physics and Astronomy(all)
- Instrumentation
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
2023 26th International Conference on Information Fusion, FUSION 2023. 2023.
Research output: Chapter in book/report/conference proceeding › Conference contribution › Research › peer review
}
TY - GEN
T1 - Improved Multi-Scale Grid Rendering of Point Clouds for Radar Object Detection Networks
AU - Köhler, Daniel
AU - Quach, Maurice
AU - Ulrich, Michael
AU - Meinl, Frank
AU - Bischoff, Bastian
AU - Blume, Holger
N1 - This work was supported by the German Federal Ministry of Education and Research, project ZuSE-KI-AVF under grant no. 16ME0062.
PY - 2023
Y1 - 2023
N2 - Architectures that first convert point clouds to a grid representation and then apply convolutional neural networks achieve good performance for radar-based object detection. However, the transfer from irregular point cloud data to a dense grid structure is often associated with a loss of information, due to the discretization and aggregation of points. In this paper, we propose a novel architecture, multi-scale KPPillarsBEV, that aims to mitigate the negative effects of grid rendering. Specifically, we propose a novel grid rendering method, KPBEV, which leverages the descriptive power of kernel point convolutions to improve the encoding of local point cloud contexts during grid rendering. In addition, we propose a general multi-scale grid rendering formulation to incorporate multi-scale feature maps into convolutional backbones of detection networks with arbitrary grid rendering methods. We perform extensive experiments on the nuScenes dataset and evaluate the methods in terms of detection performance and computational complexity. The proposed multi-scale KPPillarsBEV architecture outperforms the baseline by 5.37% and the previous state of the art by 2.88% in Car AP4.0 (average precision for a matching threshold of 4 meters) on the nuScenes validation set. Moreover, the proposed single-scale KPBEV grid rendering improves the Car AP4.0 by 2.90% over the baseline while maintaining the same inference speed.
AB - Architectures that first convert point clouds to a grid representation and then apply convolutional neural networks achieve good performance for radar-based object detection. However, the transfer from irregular point cloud data to a dense grid structure is often associated with a loss of information, due to the discretization and aggregation of points. In this paper, we propose a novel architecture, multi-scale KPPillarsBEV, that aims to mitigate the negative effects of grid rendering. Specifically, we propose a novel grid rendering method, KPBEV, which leverages the descriptive power of kernel point convolutions to improve the encoding of local point cloud contexts during grid rendering. In addition, we propose a general multi-scale grid rendering formulation to incorporate multi-scale feature maps into convolutional backbones of detection networks with arbitrary grid rendering methods. We perform extensive experiments on the nuScenes dataset and evaluate the methods in terms of detection performance and computational complexity. The proposed multi-scale KPPillarsBEV architecture outperforms the baseline by 5.37% and the previous state of the art by 2.88% in Car AP4.0 (average precision for a matching threshold of 4 meters) on the nuScenes validation set. Moreover, the proposed single-scale KPBEV grid rendering improves the Car AP4.0 by 2.90% over the baseline while maintaining the same inference speed.
KW - cs.CV
KW - cs.AI
KW - cs.LG
KW - cs.RO
UR - http://www.scopus.com/inward/record.url?scp=85171583466&partnerID=8YFLogxK
U2 - 10.48550/arXiv.2305.15836
DO - 10.48550/arXiv.2305.15836
M3 - Conference contribution
SN - 979-8-3503-1320-8
BT - 2023 26th International Conference on Information Fusion, FUSION 2023
T2 - 26th International Conference on Information Fusion, FUSION 2023
Y2 - 27 June 2023 through 30 June 2023
ER -