Application Of Cerebral Hemorrhage CT Image Segmentation Based On DICOM Image

Along with science and technology progress, people already could obtain the massive high clear medical images through each advanced medicine instrument. The CT image is highly by its resolution, low cost, mature technology and so on.it become the major way in inspection and the pathology research. However, CT image has contens massive information, and is accompanied by the noise information frequently, to a great extent to doctor's individual experience. For to organizes the organ as well as the infection region makes the direct-viewing quota survey, to enhance the CT image the readability. Carries on computer auxiliary processing to the gathering medicine CT image to have the very vital significance. The image division is in the imagery processing important research direction, is playing the strong character in the medicine imagery processing.The former people extraction frequently the DICOM document's pictorial information, transforms to the bitmap document,then carry on the further imagery processing work. But this change process has the very obvious shortcoming, The DICOM phantom's picture element spot which we obtains from the CT instrument is with 12 or 16 binary messages expressed that grey level from-1000 to 3000 between. But the BMP form image's grey level only used eight binary numbers to demonstrate. Will therefore lose the image in this conversion process the massive information. The present paper take DICOM the form skull CT image as the object of study, first introduced cerebral hemorrhage's cause of disease as well as the clinical manifestation briefly. Afterward introduced the CT instrument's principle of work and the CT image's basic format, this includes the CT image the gain, the CT fault scanning type, the DICOM form medicine image read as well as the image display packing.In the third chapter, we first briefly introduced image division and so on picture element, neighborhood, grey level some basic concepts. Afterward introduced the main algorithm of current image segmentation and the trend of development with emphasis. Finally introduced with emphasis standard FCM and the improvement algorithm's principle and the inferential reasoning process, draw out this article the algorithm.This article divides into two steps to carry on segment processing to the image. In order to eliminate outside the skull and the skull the partial information disturbance, this article first selects the method which the threshold value and the linear trace unify to the image to carry on the pre-division. Obtains the complete encephalic part. The second step uses the improvement the FCM algorithm to carry on the thin division to the image, the extraction encephalic part. In this step, to enhance the algorithm the anti-chirp nature simultaneously maintains the algorithm the operating efficiency, this article increases in the FCM objective function has taken into consideration the neighborhood information the penalty item and the control template. Compares with the former algorithm, after the improvement algorithm had the distinct improvement in the anti-chirp natural and the running rate aspect. Finally this article uses two time cluster the methods after dividing the image to carry on the infection information the extraction, the survey, obtains in each CT image infection's area, then extracts the entire infection's volume.