A Hybrid Approach to MR Imaging Segmentation Using Unsupervised Clustering and Approximate Reducts