Automated brain tumor segmentation on MR images based on neutrosophic set approach

Brain tumor segmentation for MR images is a difficult and challenging task due to variation in type, size, location and shape of tumors. This paper presents an efficient and fully automatic brain tumor segmentation technique. This proposed technique includes non local preprocessing, fuzzy intensification to enhance the quality of the MR images, k-means clustering method for brain tumor segmentation. The results are evaluated based on accuracy, sensitivity, specificity, false positive rate, false negative rate, Jaccard similarity metric and Dice coefficient. The preliminary results show 100% detection rate in all 20 test sets.