Linear Gaussian Bounding Box Representation and Ring-Shaped Rotated Convolution for Oriented Object Detection

TL;DR

This paper introduces a linear Gaussian bounding box (LGBB) representation and ring-shaped rotated convolution (RRC) to improve oriented object detection by enhancing boundary continuity, numerical stability, and feature aggregation.

Contribution

The paper proposes LGBB for stable, boundary-continuous bounding box representation and RRC for efficient rotation-sensitive feature extraction, advancing oriented object detection.

Findings

LGBB outperforms existing OBB representations in stability and boundary continuity.

RRC improves feature aggregation speed and rotation sensitivity.

Combined LGBB and RRC achieve state-of-the-art detection accuracy.

Abstract

In oriented object detection, current representations of oriented bounding boxes (OBBs) often suffer from boundary discontinuity problem. Methods of designing continuous regression losses do not essentially solve this problem. Although Gaussian bounding box (GBB) representation avoids this problem, directly regressing GBB is susceptible to numerical instability. We propose linear GBB (LGBB), a novel OBB representation. By linearly transforming the elements of GBB, LGBB avoids the boundary discontinuity problem and has high numerical stability. In addition, existing convolution-based rotation-sensitive feature extraction methods only have local receptive fields, resulting in slow feature aggregation. We propose ring-shaped rotated convolution (RRC), which adaptively rotates feature maps to arbitrary orientations to extract rotation-sensitive features under a ring-shaped receptive field, rapidly aggregating features and contextual information. Experimental results demonstrate that LGBB and RRC achieve state-of-the-art performance. Furthermore, integrating LGBB and RRC into various models effectively improves detection accuracy.