Alpha-IoU: A Family of Power Intersection over Union Losses for Bounding Box Regression

• DOI: 10.48550/arxiv.2110.13675
• Publication Date: 2021-01-01
• Authors: He, Jiabo; Erfani, Sarah; Ma, Xingjun; Bailey, James; Chi, Ying; Hua, Xian-Sheng
• Keywords: FOS: Computer and information sciences; Computer Vision and Pattern Recognition (cs.CV); Computer Science - Computer Vision and Pattern Recognition; 0202 electrical engineering, electronic engineering, information engineering; 02 engineering and technology

Bounding box (bbox) regression is a fundamental task in computer vision. So far, the most commonly used loss functions for bbox regression are the Intersection over Union (IoU) loss and its variants. In this paper, we generalize existing IoU-based losses to a new family of power IoU losses that have a power IoU term and an additional power regularization term with a single power parameter $��$. We call this new family of losses the $��$-IoU losses and analyze properties such as order preservingness and loss/gradient reweighting. Experiments on multiple object detection benchmarks and models demonstrate that $��$-IoU losses, 1) can surpass existing IoU-based losses by a noticeable performance margin; 2) offer detectors more flexibility in achieving different levels of bbox regression accuracy by modulating $��$; and 3) are more robust to small datasets and noisy bboxes.