TY - GEN
T1 - HLA
T2 - 2022 International Conference on Automation, Robotics and Computer Engineering, ICARCE 2022
AU - Chen, Qimeng
AU - Zheng, Tong
AU - Liu, Liu
AU - Yu, Longji
AU - Chen, Zhong
N1 - Publisher Copyright:
© 2022 IEEE.
PY - 2022
Y1 - 2022
N2 - The existing state-of-the-arts two-stage oriented object detectors have no significant improvement in the label assignment strategies, and the most widely-used one is the so-called Max IoU Assigner (MIA). In this paper, we first illustrate that MIA may cause matching conflicts in some cases, hinder the matching of ground-truth (GT) boxes with high-quality samples, which is extremely harmful to the training process. After that, we propose a Harmonized Label Assigner (HLA) for the oriented RPN, which can automatically harmonize the assignment priority of each GT box according to the corresponding number of candidate samples, solve the matching conflicts, and improve the detection accuracy of the two-stage oriented detectors. Finally, we implement the proposed HLA on Oriented R-CNN and conduct sufficient experiments on two public datasets (MAR20 and HRSC2016). Without tricks, our HLA significantly improves the detection accuracy of the detector to 83.97% mAP (on MAR20) and 90.42% mAP (on HRSC2016), respectively.
AB - The existing state-of-the-arts two-stage oriented object detectors have no significant improvement in the label assignment strategies, and the most widely-used one is the so-called Max IoU Assigner (MIA). In this paper, we first illustrate that MIA may cause matching conflicts in some cases, hinder the matching of ground-truth (GT) boxes with high-quality samples, which is extremely harmful to the training process. After that, we propose a Harmonized Label Assigner (HLA) for the oriented RPN, which can automatically harmonize the assignment priority of each GT box according to the corresponding number of candidate samples, solve the matching conflicts, and improve the detection accuracy of the two-stage oriented detectors. Finally, we implement the proposed HLA on Oriented R-CNN and conduct sufficient experiments on two public datasets (MAR20 and HRSC2016). Without tricks, our HLA significantly improves the detection accuracy of the detector to 83.97% mAP (on MAR20) and 90.42% mAP (on HRSC2016), respectively.
KW - label assignment
KW - oriented object detection
KW - two-stage
UR - https://www.scopus.com/pages/publications/85149433627
U2 - 10.1109/ICARCE55724.2022.10046644
DO - 10.1109/ICARCE55724.2022.10046644
M3 - 会议稿件
AN - SCOPUS:85149433627
T3 - 2022 International Conference on Automation, Robotics and Computer Engineering, ICARCE 2022
BT - 2022 International Conference on Automation, Robotics and Computer Engineering, ICARCE 2022
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 16 December 2022 through 17 December 2022
ER -