# Deep learning for cardiac image segmentation: A review

**Authors:** Chen Chen, Chen Qin, Huaqi Qiu, Giacomo Tarroni, Jinming Duan, Wenjia Bai, Daniel Rueckert  
**arXiv:** [1911.03723](https://arxiv.org/abs/1911.03723)  
**Date:** 2019-11-09  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/1911.03723v1/

## Key points

- Deep learning has become the most widely used approach for cardiac image segmentation
- The paper provides a comprehensive review of over 100 papers on deep learning techniques for cardiac image segmentation
- Various imaging modalities such as MRI, CT, and US are covered in the review
- Major anatomical structures of interest including ventricles, atria, and vessels are discussed
- The paper includes a summary of publicly available cardiac image datasets and code repositories to promote reproducible research
- Challenges and limitations of current deep learning-based approaches in cardiac image segmentation are discussed, including scarcity of labeled data, generalizability across domains, and interpretability issues
- Potential directions for future research are suggested to address these challenges
- The paper serves as a valuable resource for researchers in the field by providing an extensive review while highlighting areas that require further investigation

## Summary

In recent years, deep learning has emerged as the most widely used approach for cardiac image segmentation. In this paper titled "Deep learning for cardiac image segmentation: A review," authors Chen Chen, Chen Qin, Huaqi Qiu, Giacomo Tarroni, Jinming Duan, Wenjia Bai and Daniel Rueckert provide a comprehensive review of over 100 papers that utilize deep learning techniques for cardiac image segmentation. The review covers various imaging modalities commonly used in cardiac imaging such as magnetic resonance imaging (MRI), computed tomography (CT) and ultrasound (US). Furthermore, the authors focus on major anatomical structures of interest including ventricles, atria and vessels. One notable aspect of this paper is the inclusion of a summary of publicly available cardiac image datasets and code repositories. This addition aims to promote reproducible research by providing researchers with access to relevant data and code resources. Additionally, the authors discuss the challenges and limitations associated with current deep learning-based approaches in cardiac image segmentation such as scarcity of labeled data for training models; generalizability of models across different domains; and interpretability issues. To address these challenges they suggest potential directions for future research. Overall, this paper serves as a valuable resource for researchers in the field of cardiac image segmentation using deep learning techniques. It provides an extensive review of existing literature while highlighting areas that require further investigation. By addressing key challenges and suggesting future research directions it contributes to advancing the field towards more accurate and reliable cardiac image segmentation methods.