Automatic Segmentation Technique For Detection of Myocardial Contours in Cardiac MRI
Sushmitha R, Dr Arbind Kumar Gupta, Chandrakala B M; 4th international conference on communications and electronic systems (ICCES-2019) 17-19 Jul 2019
Segmentation of myocardial contours serves a prerequisite for viability assessment of myocardium which is essential for patient’s diagnosis and treatment planning. However, segmenting the myocardial contours more precisely is quite a challenging task. This paper presents an automated segmentation technique for detection of myocardial contours in cardiac MRI frames. Initially, a geometrical template is initialized on each frame using the co-relation technique to obtain the myocardium region. Later a curved multi-planner reconstruction (curved MPR) image is constructed to detect the endocardial and epicardial contours. Automated segmentation is tested on 12 MRI frames and accuracy is validated on three frames with respect to manually traced endo-epicardial boundaries.