Research Of Body-Fat Segmentation Algorithm Based On CT Image And Realization Of Computer-Aided Testing System

With the improvement of living level, more and more people fall obesity, a kind of chronic diseases, which has lots of complications, such as metabolic disorders, diabetes, hypertension, coronary heart disease, cardiovascular disease and so on. The content and distribution of body fat have been suggested as playing an important and etiologic role in the diagnoses of obesity. Lots of techniques had been used to do this job, but all of them cannot work fully automatically and exactly.The main goal of this paper is to develop digital image processing, pattern recognition, and morphology image processing methods to design a computer-aided body fat testing system. With the images gotten from Computerized Tomography, this paper puts emphasis on the research of image segmentation in space domain, extraction and quantization body fat. In order to find an efficient method to test the content and distribution of body fat and diagnose the obesity noninvasive.The research work of this paper includes the following five respects:Three projects were studied out to capture CT image. Main technical parameters, error sources, the merits and demerits were analyzed to every kind of projects. Capturing CT images were implemented successfully.In the research of segmentation, three calculating methods were developed, such as: the automatic threshold segmentation algorithm, the threshold segmentation algorithm modified by morphological reconstruction, and the potential function clustering segmentation algorithm.Under the environment of MATLAB, we have accomplished and assessed the data processing algorithm, and designed the GUI to make users easily control. We have also developed a corresponding software package under the environment of VC++ 6.0 named computer-aided body fat testing system, and which has a user-friendly interface with Windows style.A fat content calculation algorithm was designed based on CT image segmentation. We quantified every algorithm using many groups of CT images. After personally and impersonally validated, the results show that computer-aided body fat testing system can meet the clinical requests.