An improved method of image segmentation using fuzzy c-means

We propose a new method by incorporating improved k-means and modified fuzzy c-means clustering techniques for segmenting medical images. Segmentation of medical images plays a key role in estimating the object boundary and abnormalities, if any. Moreover, it is an important process to extract more information from complex medical images. Magnetic resonance images may contain noise due to the operator performance and the interference of radio frequency coil; hence it makes a challenging problem for segmenting such images in medical field. Fuzzy C-Means (FCM) is a flexible approach for the automatic segmentation of medical images. K-means has low computational complexity.