HLA: Harmonized Label Assigner for Two-stage Oriented Object Detection

  • Qimeng Chen*
  • , Tong Zheng
  • , Liu Liu
  • , Longji Yu
  • , Zhong Chen
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

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.

Original languageEnglish
Title of host publication2022 International Conference on Automation, Robotics and Computer Engineering, ICARCE 2022
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665475488
DOIs
StatePublished - 2022
Externally publishedYes
Event2022 International Conference on Automation, Robotics and Computer Engineering, ICARCE 2022 - Virtual, Online, China
Duration: 16 Dec 202217 Dec 2022

Publication series

Name2022 International Conference on Automation, Robotics and Computer Engineering, ICARCE 2022

Conference

Conference2022 International Conference on Automation, Robotics and Computer Engineering, ICARCE 2022
Country/TerritoryChina
CityVirtual, Online
Period16/12/2217/12/22

Keywords

  • label assignment
  • oriented object detection
  • two-stage

Fingerprint

Dive into the research topics of 'HLA: Harmonized Label Assigner for Two-stage Oriented Object Detection'. Together they form a unique fingerprint.

Cite this