# Generalized Intersection over Union: A Metric and A Loss for Bounding Box Regression

## TL;DR

This paper introduces a generalized IoU (GIoU) metric and loss function to improve bounding box regression in object detection, addressing limitations of traditional IoU and enhancing performance on benchmarks.

## Contribution

The paper proposes GIoU as a new metric and loss function that directly optimizes bounding box overlap, improving object detection accuracy.

## Key findings

- GIoU outperforms IoU in bounding box regression tasks.
- Incorporating GIoU improves detection performance on PASCAL VOC and MS COCO.
- GIoU provides a more consistent gradient for non-overlapping boxes.

## Abstract

Intersection over Union (IoU) is the most popular evaluation metric used in the object detection benchmarks. However, there is a gap between optimizing the commonly used distance losses for regressing the parameters of a bounding box and maximizing this metric value. The optimal objective for a metric is the metric itself. In the case of axis-aligned 2D bounding boxes, it can be shown that $IoU$ can be directly used as a regression loss. However, $IoU$ has a plateau making it infeasible to optimize in the case of non-overlapping bounding boxes. In this paper, we address the weaknesses of $IoU$ by introducing a generalized version as both a new loss and a new metric. By incorporating this generalized $IoU$ ($GIoU$) as a loss into the state-of-the art object detection frameworks, we show a consistent improvement on their performance using both the standard, $IoU$ based, and new, $GIoU$ based, performance measures on popular object detection benchmarks such as PASCAL VOC and MS COCO.