Segmentation of Brain Portion From MRI of Head Scans Using K-Means Cluster
K. Somasundaram, S. Vijayalakshmi, Thiruvenkadam Kalaiselvi · 2011
Segmentation is one of the essential processes in i mage processing. The objective of this paper is to segment human brain area from other non-brain area in MRI of head scans. In our method, brain portion is detected us ing clustering technique. The proposed method clusters the image i nto 15 regions using k-mean clustering technique. T he non brain parts like skull, sclera, fat, skin were clustered to regions according to its intensity. These region s were eliminated and the remaining regions are merged to form the brain portion. The proposed model has been tested on vari ous image slices and found to give good segmentation. Experimental results show that the proposed method gives a n average value of 0.95 for Jaccard co-efficient, 0.97 for Dice, 0. 96 for Sensitivity and 0.98 for specificity.