Auto Segmentation Method for Electromagnetic Radiation Camera Myocardial Images

Abstract

Electromagnetic radiation camera (ERC) provides functional information about the left ventricle. We need new methods of quantitative analysis, such as the segmentation myocardial images for maximum utilization of ERC. The level set method(semi-automatic) based approach method that has excellent potential with regarding mapping topological change in 2D images but the crucial point is when stopping iteration to obtain the optimal segmentation result. This study suggests the Courant–Friedrichs–Lewy (CFL) condition, feature measurement index (FMI), and modified Thompson Tau technique (MTTT) to investigate the optimal convergence problem. The CFL condition contributes to the stable numerical iteration of the level set, whereas the FMI and MTTT are used for exact iteration stopping. This study compared the auto segmentation method and the manual segmentation method through assessments. The results were so matched (82.39±13.92 %) so we can develop a 3D model that could be used to diagnose ischemic heart disease and infarction. The proposed segmentation method can help to diagnose the disease conveniently and accurately in the clinic.