CLASSIFICATION OF MRI BRAIN IMAGES USING DISCRETE WAVELET TRANSFORM AND K - NN

Presented work is a feature extraction and classification study for diagnosis of Brain cancer (abnormal) and normal brain images. The proposed method consists of two stages, namely feature extraction and classification. In the feature extraction stage features are extracted using discrete wavelet transformation (DWT) from MRI images. In the second stage, a non - parametric statist ic technique based on k - nearest neighbor (k - NN) algorithm is used for classification. The classifier has been used to classify images as normal or abnormal MRI brain images. We applied this method on 80 images (50 training images divided into 25 normal, 25 abnormal) and (30 test images divided into 15 normal, 15 abnormal) and dimensions of the images 256*256 pixel. The classification accuracies on both training and test images are 98%, has been obtained by the proposed classifier k - NN.