The KFIoU Loss for Rotated Object Detection

TL;DR

This paper introduces the KFIoU loss, an efficient and differentiable approximation for SkewIoU in rotated object detection, improving training stability and accuracy across 2D and 3D datasets.

Contribution

The paper proposes a novel Gaussian-based approximate SkewIoU loss (KFIoU) that simplifies implementation and enhances performance over existing methods, including 3D extension.

Findings

KFIoU outperforms exact SkewIoU loss in experiments

Effective in 2D and 3D rotated object detection

Works well across various datasets and detectors

Abstract

Differing from the well-developed horizontal object detection area whereby the computing-friendly IoU based loss is readily adopted and well fits with the detection metrics. In contrast, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. In this paper, we propose an effective approximate SkewIoU loss based on Gaussian modeling and Gaussian product, which mainly consists of two items. The first term is a scale-insensitive center point loss, which is used to quickly narrow the distance between the center points of the two bounding boxes. In the distance-independent second term, the product of the Gaussian distributions is adopted to inherently mimic the mechanism of SkewIoU by its definition, and show its alignment with the SkewIoU loss at trend-level within a certain distance (i.e. within 9 pixels). This is in contrast to recent Gaussian modeling based rotation detectors e.g. GWD loss and KLD loss that involve a human-specified distribution distance metric which require additional hyperparameter tuning that vary across datasets and detectors. The resulting new loss called KFIoU loss is easier to implement and works better compared with exact SkewIoU loss, thanks to its full differentiability and ability to handle the non-overlapping cases. We further extend our technique to the 3-D case which also suffers from the same issues as 2-D. Extensive results on various public datasets (2-D/3-D, aerial/text/face images) with different base detectors show the effectiveness of our approach.