Segmentation of Short Axis CMR Images Using Hybrid Method

Background: Highly advanced and sophisticated imaging modality, Cardiac Magnetic Resonance (CMR) images are referred to examine the cardiac morphology and its function. Methods: In this work, the main aim is to develop a hybrid segmentation method for automatic segmentation of both left, right ventricles from short axis CMR images. In the proposed hybrid segmentation method, Fast Adaptive K-Means (FAKM) clustering method is used to locate the ventricles which are further segmented by Distance Regularized Level Set Evolution (DRLSE) method. Results: The validation parameters show that the segmentation by proposed hybrid method is better than hybrid methods like Gaussian mixture model with dynamic programming and semi-automatic method. Discussions: Further, FAKM hybrid method is evaluated based on End Systolic Volume (ESV), End Diastolic Volume (EDV) and Ejection Fraction (EF). Conclusion: The analytical result shows that the hybrid method of FAKM with DRLSE gives faster and better results.