European Journal of Biomedical and Pharmaceutical Sciences

AUTOMATIC DETECTION OF ISCHEMIC STROKE LESION USING TEXTURAL ANALYSIS FROM BRAIN CT IMAGES

Asmita Ray and *Prof. Samir Kumar Bandyopadhyay

ABSTRACT

Stroke lesion quantifying and characterizing is still an open challenge. This method diagnosis the stroke lesion accurately which helps the radiologist fast and proper treatment so that death rate due to brain stroke can be reduced. The method contains three main modules: preprocessing, segmentation and feature extraction. Pre-processing module removes noise and artifacts and abnormal region is extracted in segmentation module. The image is subdivided into four regions to find the region that has the possibility for inclusion of abnormal area. The abnormal portion which contains stroke lesion is characterized with statistical features using first-order histogram. The paper proposes a self-constructed, effective algorithm to support ischemic stroke lesion detection and feature extraction based on CT images.

Keywords: Ischemic Stroke, Computed Tomography (CT), Brain Segmentation, Statistical Features, Infarct.