Abstract

Abstract: Comparative analysis of different machine learning techniques like support vector machine (SVM), k-nearestneighbor (KNN) & Artificial neural network (ANN) are proposed to be applied for detection and classification ofcancerous and non-cancerous brain MRI images. In pre-processing for removing noise, anisotropic diffusion filtering isused and k-mean clustering based segmentation technique is used. To extract the features of segmented images discretewavelet transform (DWT) coefficient gray level co-occurrence matrix (GLCM) method is used. Accuracy of differentproposed techniques are calculated with the help of confusion matrix and compared to get best classification techniquefor brain MRI images. For analysis of proposed method brain MRI classification dataset is used available at kaggle.com.Keywords: MRI Images, SVM, KNN, ANN, Segmentation, Feature Extraction, Brain MRI Image classification.