Research On Segmentation Algorithm Of Cerebral Computed Tomography Image

With the development of science and technology, medical images have become important assistant means of diagnosis and therapy. Computed Tomography (CT) image is an important method to research of pathology and anatomy, as its high resolution and little harm to human beings. However, there are many inevitable defects in medical images. In order to improve the read ability of medical image and make the doctor to adopt more effective observation and diagnoses way to anatomy structure and pathological part of the patient, it's necessary to study the medical image processing. The image segmentation technology plays an important part in image processing and therefore many researchers have paid much attention on it. The basic idea of image segmentation is to divide the image space into some special regions. As for cerebral CT image, it is to extract meaningful or useful features, which are original information in image space.Firstly, this dissertation studies the interrelated domains of cerebral Computed Tomography image segmentation, including research status quo, research significances, formalized definition, developing trends, Computed Tomography, etc, which builds a systematic theory framework of studies on cerebral Computed Tomography image segmentation. Secondly, under the direct of these theories discussed above, this dissertation analyses some the existing techniques for image segmentation, and that proposes two new methods for cerebral Computed Tomography image. At last, we summarize this paper, and propose the research plan for next step.A novel algorithm is proposed in this paper in terms of the existing extraction algorithms for intracranial structures on cerebral Computed Tomography lacking automation. The skull image from CT is characterized by great width and high gray-level, considering that, the proposed algorithm uses linear filtering to extract the outline of skull. And then, mathematic morphology and horizontal scanning algorithm is employed for automatic segmentation of intracranial structures on cerebral CT. Experiment on 100 cases of cerebral CT demonstrates that the ratio of automatic segmentation reaches up to 99%, and the results are accurate and encouraging.To develop a computer aided detection system that improves diagnostic accuracy of intracranial hemorrhage (IH) on brain CT. A novel method for CT image segmentation of brain is proposed, with which, several regions that are suspicious of hemorrhage can be segmented rapidly and effectively. Algorithm of extracting intracranial area was introduced firstly to extract intracranial area. Secondly, FCM was employed twice, we named it with TFCM. FCM was first employed to identify areas of intracranial hemorrhage. Finally, FCM was employed to segment the lesions. Experimental results on real medical images demonstrate the efficiency and effectiveness of the proposed algorithm.