MRI retinal image segmentation using integrated approach of fuzzy c-means clustering, and active contouring

In order to locate the different objects, shapes and structures in digital images, segmentation is the most acceptable and popular choice. This paper combines fuzzy c means clustering (ECM) and mathematical morphology (MM) for the segmentation of MRI retinal images and spinal cord Xray images and for the same set of medical images, segmentation using active contouring (Level set method) is carried out for better interpretation of region of interest. Further the results are comparatively analyzed based on visual quality as well as performance of segmentation algorithms are evaluated on the parameters like, boundary displacement error (BDE) and entropy factor. Simulation results states that the entropy is 1.3387 with integrated approach of FCM & morpholo*,. And it is 5.3127 in case of active contouring, which results in high randomness in the segmented variables. Based on the evaluation, it can be suggested that integrated approach of FCM & erosion provide the better segmentation accuracy for medical images.