Dynamic Anchor Learning for Arbitrary-Oriented Object Detection

Abstract

Arbitrary-oriented objects widely appear in natural scenes, aerial photographs, remote sensing images, etc., thus arbitrary-oriented object detection has received considerable attention. Many current rotation detectors use plenty of anchors with different orientations to achieve spatial alignment with ground truth boxes, then Intersection-over-Union (IoU) is applied to sample the positive and negative candidates for training. However, we observe that the selected positive anchors cannot always ensure accurate detections after regression, while some negative samples can achieve accurate localization. It indicates that the quality assessment of anchors through IoU is not appropriate, and this further lead to inconsistency between classification confidence and localization accuracy. In this paper, we propose a dynamic anchor learning (DAL) method, which utilizes the newly defined matching degree to comprehensively evaluate the localization potential of the anchors and carry out a more efficient label assignment process. In this way, the detector can dynamically select high-quality anchors to achieve accurate object detection, and the divergence between classification and regression will be alleviated. With the newly introduced DAL, we achieve superior detection performance for arbitrary-oriented objects with only a few horizontal preset anchors. Experimental results on three remote sensing datasets HRSC2016, DOTA, UCAS-AOD as well as a scene text dataset ICDAR 2015 show that our method achieves substantial improvement compared with the baseline model. Besides, our approach is also universal for object detection using horizontal bound box. The code and models are available at https://github.com/ming71/DAL.

arXiv commentAccepted to AAAI 2021. The code and models are available at https://github.com/ming71/DAL

References in corpus (10)

• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• R-FCN: Object Detection via Region-based Fully Convolutional Networks
• What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?
• Gliding vertex on the horizontal bounding box for multi-oriented object detection
• Automatic Ship Detection of Remote Sensing Images from Google Earth in Complex Scenes Based on Multi-Scale Rotation Dense Feature Pyramid Networks
• R2CNN: Rotational Region CNN for Orientation Robust Scene Text Detection
• FreeAnchor: Learning to Match Anchors for Visual Object Detection
• R3Det: Refined Single-Stage Detector with Feature Refinement for Rotating Object
• Consistent Optimization for Single-Shot Object Detection
• Align Deep Features for Oriented Object Detection

Cited by in corpus (6)

• Anchor-free Oriented Proposal Generator for Object Detection
• CFC-Net: A Critical Feature Capturing Network for Arbitrary-Oriented Object Detection in Remote Sensing Images
• Rethinking Rotated Object Detection with Gaussian Wasserstein Distance Loss
• Arbitrary-Oriented Ship Detection through Center-Head Point Extraction
• Learning Oriented Remote Sensing Object Detection via Naive Geometric Computing
• MidNet: An Anchor-and-Angle-Free Detector for Oriented Ship Detection in Aerial Images