A cascade approach for automatic segmentation of cardiac structures in short-axis cine-MR images using deep neural networks

作者：

Highlights：

• A computational method for automatic segmentation of cardiac structures is presented.

• Our method uses a cascaded approach combining FCNs and image processing techniques.

• The proposed method is validated with cine-MR images from ACDC dataset.

• The method achieves promising results, being competitive in the ACDC challenge.

摘要

•A computational method for automatic segmentation of cardiac structures is presented.•Our method uses a cascaded approach combining FCNs and image processing techniques.•The proposed method is validated with cine-MR images from ACDC dataset.•The method achieves promising results, being competitive in the ACDC challenge.

论文关键词：Cardiac structures segmentation,Fully convolutional networks,Cine-MRI,ACDC dataset

论文评审过程：Received 18 May 2021, Revised 10 February 2022, Accepted 19 February 2022, Available online 4 March 2022, Version of Record 9 March 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.116704