Paper Title
Automatic Brain MR Image Tumor Detection using Region Growing
Abstract
Seeded region growing is very popular and attractive algorithm for image segmentation. It involves high-level
knowledge of image components in the seed selection procedure. The seed selection procedure suffers from problems of
automatic seed generation. In this work a new method for seed selection is proposed for region growing algorithm. The
proposed method uses fuzzy c means algorithm for automatic seed selection in region growing and hence segment tumor
core from MR brain images. Experiments for the proposed method are conducted on FLAIR images from publically
available BRATS database. Validation of results with ground truth demonstrates an average Accuracy of 97%, Sensitivity of
70% and Specificity of 98%.
Keywords - FCM, region growing, automatic seed selection, tumor segmetaion