OSKDet: Towards Orientation-sensitive Keypoint Localization for Rotated Object Detection

TL;DR

OSKDet introduces an orientation-sensitive keypoint detection method that enhances rotated object detection accuracy by learning shape and direction implicitly, utilizing a rotation-aware deformable convolution and a keypoint reorder algorithm.

Contribution

The paper presents a novel rotated detector, OSKDet, with orientation-sensitive heatmaps and a keypoint reorder algorithm to improve detection accuracy for rotated objects.

Findings

Achieves state-of-the-art AP of 77.81% on DOTA

Improves localization accuracy with orientation-sensitive heatmaps

Utilizes rotation-aware deformable convolution for border feature extraction

Abstract

Rotated object detection is a challenging issue of computer vision field. Loss of spatial information and confusion of parametric order have been the bottleneck for rotated detection accuracy. In this paper, we propose an orientation-sensitive keypoint based rotated detector OSKDet. We adopt a set of keypoints to characterize the target and predict the keypoint heatmap on ROI to form a rotated target. By proposing the orientation-sensitive heatmap, OSKDet could learn the shape and direction of rotated target implicitly and has stronger modeling capabilities for target representation, which improves the localization accuracy and acquires high quality detection results. To extract highly effective features at border areas, we design a rotation-aware deformable convolution module. Furthermore, we explore a new keypoint reorder algorithm and feature fusion module based on the angle distribution to eliminate the confusion of keypoint order. Experimental results on several public benchmarks show the state-of-the-art performance of OSKDet. Specifically, we achieve an AP of 77.81% on DOTA, 89.91% on HRSC2016, and 97.18% on UCAS-AOD, respectively.