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

Abstract

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 $\alpha$. We call this new family of losses the $\alpha$-IoU losses and analyze properties such as order preservingness and loss/gradient reweighting. Experiments on multiple object detection benchmarks and models demonstrate that $\alpha$-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 $\alpha$; and 3) are more robust to small datasets and noisy bboxes.

References in corpus (10)

• YOLOv4: Optimal Speed and Accuracy of Object Detection
• Deformable DETR: Deformable Transformers for End-to-End Object Detection
• MMDetection: Open MMLab Detection Toolbox and Benchmark
• Hybrid Task Cascade for Instance Segmentation
• Normalized Loss Functions for Deep Learning with Noisy Labels
• Loss Function Discovery for Object Detection via Convergence-Simulation Driven Search
• Towards Noise-resistant Object Detection with Noisy Annotations
• Auto Seg-Loss: Searching Metric Surrogates for Semantic Segmentation
• AutoLoss-Zero: Searching Loss Functions from Scratch for Generic Tasks
• Learning Non-Unique Segmentation with Reward-Penalty Dice Loss

Cited by in corpus (1)

• SpineOne: A One-Stage Detection Framework for Degenerative Discs and Vertebrae