| Summary: | 碩士 === 國立中正大學 === 資訊管理系研究所 === 105 === According to statistics of government, cerebrovascular diseases rank number three among the top ten leading causes of death in Taiwan. Intracranial hemorrhage, or hemorrhagic stroke, is a kind of cerebrovascular diseases. It may lead to hemiplegia, language disorder, unconsciousness, and even death. Moreover, the mortality rate of hemorrhagic stroke is higher than the rate of ischemic stroke.
Computed tomography is the main imaging modality used for diagnosis and evaluation of intracranial hemorrhage, which can distinguish the type of stroke and the location and the size of the hematoma. Because hematoma is of high density in computed tomography, it is easy to identify. The location and the volume of the hematoma are the basis for clinicians to decide whether to operate. Due to the irregular shape of hematoma, the volume of hematoma cannot measure accurately if only use its length, width and height to calculate.
Therefore, we propose an automatic image processing method that aims to provide a powerful tool for clinicians to evaluate the intracranial hemorrhage in computed tomography, and to reduce the time of manual selection. Combining adaptive thresholding and region growing, we perform otsu’s method to find initial seeds automatically. Through this method, we can determine the gray value of the seed point by dynamic threshold obtained from Otsu’s method. The results show that the region circled by this study and the contour manually circled by the doctor have great similarity, and yield good accuracy and specificity.
|
|---|
