Analysis of Stroke using texture features

Analyzing the occurrence of Stroke is a challenging issue among different patients. The research work presented here is a two phase classification method in which the initial phase automatically detects the stroke affected and normal Computed Tomography (CT) images while the second phase classifies the hemorrhagic and ischemic stroke from a set of stroke affected images. The proposed method has been tested with a number of CT brain images and has achieved promising results. A classification accuracy of 91% has been obtained by SVM(Support Vector Machine ). The performance evaluation of the proposed approach validates its effectiveness and robustness.