Cad Scheme For Detection Of Lacunar Infarcts In Brain Mr Image

The detection and management of asymptomatic lacunar infarcts in MRI images are an important task for radiologists to prevent the occurrence of cerebral apoplexy However, it is difficult for radiologists to identify the lacunar infarcts correctly in MRI images. Therefore, the purpose of our study was to develop a computer-aided diagnosis scheme for detection of lacunar infarcts in order to assist radiologists' interpretation as a "second opinion.'' First, we segmented the cerebral parenchymal region in T1-weighted image by using a region growing technique. For identifying the initial candidates of lacunar infarcts, the top-hat transform and multiple-phase binarization were then applied to the T2-weighted image within the segmented parenchymal region. The locations x and y, density differences from T1- and T2-weighted images, nodular components from scale 1 to 4, and nodular and linear components from scales 1 to 4 were determined in the initial candidates regions. The rule-based schemes and an artificial neural network with 12 features were applied to the initial candidates for distinguishing between lacunar infarcts and false positives (FPs). The sensitivity of the detection of lacunar infarcts was 96.8%(90/93) with 0.71 FPs per image. Our computerized scheme would be useful in assisting radiologists for identifying lacunar infarcts in MR images.