A Novel Segmentation Algorithm for Feature Extraction of Brain MRI Tumor

A new algorithm is projected in this paper for the identification and classification of tumors. For this, a set of MRI slices is considered from the database. As the images from electronic equipment contain noise, first the denoising of images is done using wavelets. Now, the identification of tumor is done by segmentation. Initially, the existing methods like expectation–maximization, histogram, and object-based thresholding are analyzed and implemented. But some of the features are missing in all these methods. So a new algorithm is proposed in which all the features from above methods are fused. The total analysis is done for 2D images, and the results obtained are in 2D. The performance analysis of the existing and proposed algorithms is compared in terms of size of the resultant tumor.