Semi Automatic Method for Basal Ganglia and White Matter Lesion Segmentation in MRI Images of Cronic Stroke Patients Using Adaptive Otsu

摘要

Ischemic stroke is a stroke that occurs because the artery blockage that arises brings blood to the brain. Image of Magnetic Resonance Imaging (MRI) is widely used in diagnosing the type of brain ischemic stroke including stroke. To detect ischemic stroke on MRI, T2 Tl sequence utilizing a san. Fluid Attenuation Inversion Recovery (Flair) is suitable for detecting brain infarct. The difficulties faced by radiology experts occurs in the interpretation of the imagery that happened in the MRI images axial flair because of the complexity of the site caused by the artifacts caused by the movement of Physiology, the fault geometry, and the picture is not irregular. This research aims to minimize the occurrence of errors of interpretation of the MRI images axial Flair caused movement of the median filter using physiology as well as perform segmentation on a Region of Interest (ROI) using adaptive segmentation of otsu ischemic lacunar infarct patients for the chronic type. This research uses data from 50 patients they would stroke ischemic lacunar infarct with image sequences used are MRI images axial Flair. The image generated by the MRI T2 axial Flair sequence on the converted to ajpeg image and put the median filter to produce a clean copy of the description of the artifact. Prior segmentation, do remove skull so that the process of segmentation, ROI can distinguish the model of white matter and gray matter. The result of the segmentation will be produced as well as lacunar ROI value accuracy, and specificity segmentation comparison image of ground truth carried out by expert radiology of 99.99% and 96.8%.