Ultrasound image processing and 3D reconstruction of heart in patients with cardiomyopathy

Abstract

This paper describes the development of a machine learning based diagnostic tool that analyzes cardiac ultrasound images of patients with cardiomyopathy from 4 chamber/2 chamber apical and M mode views. The method was implemented using a dataset containing 1809 images in apical view and 53 images in M-mode view from patients with cardiomyopathy. Comparing manually annotated LV and automatically calculated parameters,we achieve dice coefficient of 92.091% for segmentation and an average root mean square error (RMSE) of 0.3052cm for parameter extraction in apical view images and an average RMSE of 1.3548cm for parameter extraction in M-mode view images. We reconstruct a 3D model of the left ventricle using calculated parameters.