Computer-Aided Diagnosis Research Of Hypertensive Cerebral Hemorrhage CT Image

Cerebral hemorrhage quantity, midline shift and ventricles pressure are key indexes in the diagnosis of hypertensive cerebral hemorrhage. In view of the problems that cerebral hemorrhage quantity exists calculation error when chef formula is used to calculate the irregular, discontinuous and inconspicuous hemorrhagic region, smaller midline shift and less obvious ventricles pressure cause imprecise diagnostic information; doctors can not predict hematoma absorption, the brain tissues appear secondary injury, we investigate the research about computer-aided diagnosis of CT images. The research results improved the accuracy of cerebral hemorrhage quantity, midline shift and ventricles pressure, having important significance for doctors to diagnosis hypertensive cerebral hemorrhage(HCH).The main contents are as follows:1. Devised a segmented method about the irregular, discontinuous and inconspicuous hemorrhagicl region; in the view of this uneven gray-level, no obvious boundary cerebral hemorrhage CT images. Firstly, a fuzzy C-means clustering algorithm based on spatial information was employed to segment intracranial hemorrhage CT image. Secondly, distance regularized level set evolution model was initialized based on the results of the fuzzy clustering segmentation, finally, the hemorrhagic region was segmented by iterating. The new segmented algorithm considered the spatial and edged information, without initialization, and the double well potential function improved distance regularized item, improved the accuracy and efficiency in the process of segmentation.2. A new method about computer-aided calculating cerebral hemorrhage quantity based on segmented algorithm was designed; firstly, the proportion of hemorrhagic region was obtained by binarization, then proportion multiplied the area of CT image to calculate bleeding area, finally, adding all bleeding levels is hemorrhage quantity.3. Establishing prediction models of the hematoma absorption on the basis of analyzing clinical risk factors of hematoma absorption. Gathered 121 cases of HCH patients' clinical data and CT images. Calculating cerebral hemorrhage quantity, hematoma absorption rate and hematoma absorption proportion; analyzing clinical risk factors of hematoma absorption by SPSS software, establishing prediction model of the hematoma absorption rate and hematoma absorption proportion by binary logistic regression analysis.4. Measuring midline shift and ventricles pressure on the Matlab GUI platform, developing computer aided diagnosis system about HCH. Diagnosis system achieved accurate measurement of midline shift and ventricles pressure by conversion between edge scale of CT image and grid coordinate, calculating the ventricles' area by counting grids at the ventricle's region.