# Detecting Lesion Bounding Ellipses With Gaussian Proposal Networks

## TL;DR

This paper introduces Gaussian Proposal Networks (GPNs), an innovative extension of RPNs that detects elliptical lesions in CT scans by modeling them as Gaussian distributions, leading to improved localization accuracy over traditional bounding box methods.

## Contribution

GPNs represent lesions as Gaussian distributions and optimize ellipse detection by minimizing KL divergence, effectively capturing elliptical geometry in lesion detection.

## Key findings

- GPN outperforms RPN in lesion ellipse detection accuracy.
- GPN achieves lower localization error on DeepLesion dataset.
- The method is open sourced for community use.

## Abstract

Lesions characterized by computed tomography (CT) scans, are arguably often elliptical objects. However, current lesion detection systems are predominantly adopted from the popular Region Proposal Networks (RPNs) that only propose bounding boxes without fully leveraging the elliptical geometry of lesions. In this paper, we present Gaussian Proposal Networks (GPNs), a novel extension to RPNs, to detect lesion bounding ellipses. Instead of directly regressing the rotation angle of the ellipse as the common practice, GPN represents bounding ellipses as 2D Gaussian distributions on the image plain and minimizes the Kullback-Leibler (KL) divergence between the proposed Gaussian and the ground truth Gaussian for object localization. We show the KL divergence loss approximately incarnates the regression loss in the RPN framework when the rotation angle is 0. Experiments on the DeepLesion dataset show that GPN significantly outperforms RPN for lesion bounding ellipse detection thanks to lower localization error. GPN is open sourced at https://github.com/baidu-research/GPN