A New Brain Magnetic Resonance Imaging Segmentation Algorithm Based on Subtractive Clustering and Fuzzy C-Means Clustering

Human brain imaging provides reliable and fast information in clinical diagnosis, and a precise brain image allows researchers to effectively understand the lesions of the patients. Image segmentation on brain magnetic resonance images has been studied by researchers from different fields. However, the low accuracy as well as huge amount of computing results in the difficulties in processing. This paper explores the feasibility of segmenting images based on the practical recordings. A segmentation scheme based on subtractive clustering and fuzzy C-means clustering algorithm is developed to efficiently compensate the complexity and instability in image segmentation, and it enables exact obtaining target object from the samples by applying FCM-based clustering methods. Meanwhile, the application of median filtering removes the noise of raw images. Performance of the solution is verified based on experimental outcomes. The result shows SFCM is useful in extracting valuable information of image, which makes it a promising basis for the application of brain magnetic resonance image segmentation.