Divide and Conquer: A Large-Scale Dataset and Model for Left-Right Breast MRI Segmentation

TL;DR

This paper presents the first large-scale publicly available breast MRI dataset with explicit left-right segmentation labels and a deep-learning model trained for this task, advancing tools for women's health analysis.

Contribution

It introduces a novel, extensive dataset and a robust segmentation model, filling a critical gap in breast MRI analysis resources.

Findings

Over 13,000 annotated cases in the dataset

A deep-learning model achieving accurate left-right breast segmentation

Public availability of dataset and model for research community

Abstract

We introduce the first publicly available breast MRI dataset with explicit left and right breast segmentation labels, encompassing more than 13,000 annotated cases. Alongside this dataset, we provide a robust deep-learning model trained for left-right breast segmentation. This work addresses a critical gap in breast MRI analysis and offers a valuable resource for the development of advanced tools in women's health. The dataset and trained model are publicly available at: www.github.com/MIC-DKFZ/BreastDivider