Brain stroke detection using K-Nearest Neighbor and Minimum Mean Distance technique

This work aims to evaluate the relative performance of K-Nearest Neighbor Classifier and Minimum Mean Distance classifier of the brain stroke images. And it is a fully automated method to identify and classify an irregularity (hemorrhage) of stroke in brain. Whenever blood supply to the brain is stopped brain stroke occurs. Automatic detection and classification of MRI images brain stroke and non-stroke categories is complex phenomenon requires high level processing. In this paper the authors have proposed novel algorithm employing LabVIEW software and estimated the Identification score and Classification score and also the stroke area. Identification score for KNN method is greater than the Minimum mean distance. Both in KNN and MMD Maximum metric provides the high identification score than the Euclidian and Sum ( Manhattan metric).