Balanced Sample Assignment and Objective for Single-Model Multi-Class 3D Object Detection
IEEE Transactions on Circuits and Systems for Video Technology · 2023-09 · 33 (9) · 5036-5048
DOI: 10.1109/tcsvt.2023.3248656 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Circuits and Systems for Video Technology, 2023-09. Verfasst von Weiping Xiao und 5 weiteren Autor:innen. Die Arbeit wurde bisher 42-mal zitiert (Crossref) und nennt 43 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Ministry of Science and Technology of the People’s Republic of China.
Abstract (Originalsprache)
Accurately detecting multi-class objects in a single pass is critical but challenging for real-world autonomous driving scenarios. Several single-class anchor-based methods have recently achieved the state-of-the-art performance in the car category, but when extending to multi-class detection tasks, their performance on small objects (i.e., pedestrians and cyclists) is limited. We find that the core problem that causes this phenomenon lies in the unbalanced sample quality and the classification objective. To address this problem, we proposed a single-model multi-class 3D object detector with balanced sample assignment and objective, named BSAODet. Specifically, the quality-balanced sample assignment (QBSA) is introduced to dynamically collect stable high-quality samples for each class according to the predicted sample performance and geometric constraints. In conjunction with the QBSA, the class-balanced classification objective (CBCO) performs instance-wise label normalization and weighting on positive samples, preventing the model from biasing toward objects with more samples. Extensive experiments on the popular KITTI dataset, the latest large-scale ONCE dataset, and the challenging Waymo Open Dataset show that our method steadily improves the performance of current state-of-the-art detectors by 2–7 mAP in pedestrians and cyclists while maintaining competitiveness in cars. Moreover, our best model achieves 66.31 mAP on three classes, outperforming all published LiDAR-only detectors on the KITTI benchmark.
Förderung
- National Natural Science Foundation of China
- Ministry of Science and Technology of the People’s Republic of China
Zitieren
APA
Xiao, W., Peng, Y., Liu, C., Gao, J., Wu, Y., & Li, X. (2023). Balanced Sample Assignment and Objective for Single-Model Multi-Class 3D Object Detection. IEEE Transactions on Circuits and Systems for Video Technology, 33(9), 5036-5048. https://doi.org/10.1109/tcsvt.2023.3248656
Vancouver
Xiao W, Peng Y, Liu C, Gao J, Wu Y, Li X. Balanced Sample Assignment and Objective for Single-Model Multi-Class 3D Object Detection. IEEE Transactions on Circuits and Systems for Video Technology. 2023;33(9):5036-5048. doi:10.1109/tcsvt.2023.3248656
BibTeX
@article{Xiao2023,
author = {Xiao, Weiping and Peng, Yan and Liu, Chang and Gao, Jiantao and Wu, Yiqiang and Li, Xiaomao},
title = {Balanced Sample Assignment and Objective for Single-Model Multi-Class 3D Object Detection},
journal = {IEEE Transactions on Circuits and Systems for Video Technology},
year = {2023},
volume = {33},
number = {9},
pages = {5036--5048},
doi = {10.1109/tcsvt.2023.3248656}
}
EUVIMED