Abstract:Existing rotated object detectors are mostly inherited from the horizontal detection paradigm, as the latter has evolved into a well-developed area. However, these detectors are difficult to perform prominently in high-precision detection due to the limitation of current regression loss design, especially for objects with large aspect ratios. Taking the perspective that horizontal detection is a special case for rotated object detection, in this paper, we are motivated to change the design of rotation regression loss from induction paradigm to deduction methodology, in terms of the relation between rotation and horizontal detection. We show that one essential challenge is how to modulate the coupled parameters in the rotation regression loss, as such the estimated parameters can influence to each other during the dynamic joint optimization, in an adaptive and synergetic way. Specifically, we first convert the rotated bounding box into a 2-D Gaussian distribution, and then calculate the Kullback-Leibler Divergence (KLD) between the Gaussian distributions as the regression loss. By analyzing the gradient of each parameter, we show that KLD (and its derivatives) can dynamically adjust the parameter gradients according to the characteristics of the object. For instance, it will adjust the importance (gradient weight) of the angle parameter according to the aspect ratio. This mechanism can be vital for high-precision detection as a slight angle error would cause a serious accuracy drop for large aspect ratios objects. More importantly, we have proved that KLD is scale invariant. We further show that the KLD loss can be degenerated into the popular l(n)-norm loss for horizontal detection. Experimental results on seven datasets using different detectors show its consistent superiority, and codes are available at https://github.com/yangxue0827/RotationDetection.

Learning Modulated Loss for Rotated Object Detection

RSDet++: Point-based Modulated Loss for More Accurate Rotated Object Detection

Rotated Object Detection via Scale-Invariant Mahalanobis Distance in Aerial Images

Learning High-Precision Bounding Box for Rotated Object Detection Via Kullback-Leibler Divergence

Learning power Gaussian modeling loss for dense rotated object detection in remote sensing images

Adaptive Angle Module and Radian Regression Method for Rotated Object Detection

Rethinking Rotated Object Detection with Gaussian Wasserstein Distance Loss

The KFIoU Loss for Rotated Object Detection

FPDIoU Loss: A Loss Function for Efficient Bounding Box Regression of Rotated Object Detection

Rotating object detection in remote-sensing environment

Detecting Rotated Objects as Gaussian Distributions and its 3-D Generalization

YOLOv7-Aerial: Rotated Object Detection of Aerial Images Based on Improved YOLOv7

Point RCNN: An Angle-Free Framework for Rotated Object Detection

OSKDet: Orientation-sensitive Keypoint Localization for Rotated Object Detection

Rotated YOLOv4 with Attention-wise Object Detectors in Aerial Images

SCRDet: Towards More Robust Detection for Small, Cluttered and Rotated Objects

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

Precise Orientation Estimation for Rotated Object Detection Based on a Unit Vector Coding Approach

SCRDet++: Detecting Small, Cluttered and Rotated Objects via Instance-Level Feature Denoising and Rotation Loss Smoothing

RiDOP: A Rotation-Invariant Detector with Simple Oriented Proposals in Remote Sensing Images

Knowledge Combination to Learn Rotated Detection Without Rotated Annotation