Automatic segmentation of left and right ventricles in cardiac MRI using 3D-ASM and deep learning

作者：

Highlights：

• Manual 2D contours are used to build distance maps to get 3D ground truth shape.

• Right ventricle points from manual and deep learning are employed to rotate the coarse initial shape for a refined bi-ventricle initial estimate.

• Segmentation results from deep learning are utilized to build distance maps for the 3D-ASM matching process.

摘要

•Manual 2D contours are used to build distance maps to get 3D ground truth shape.•Right ventricle points from manual and deep learning are employed to rotate the coarse initial shape for a refined bi-ventricle initial estimate.•Segmentation results from deep learning are utilized to build distance maps for the 3D-ASM matching process.

论文关键词：Left and right ventricle segmentation,automatic initialisation,deep learning,statistical shape models

论文评审过程：Received 8 August 2020, Revised 18 April 2021, Accepted 29 April 2021, Available online 3 May 2021, Version of Record 25 May 2021.

论文官网地址：https://doi.org/10.1016/j.image.2021.116303