Mr Image Processing In The Human Body Composition Analysis

Human Body Composition Analysis (HBCA) needs to quantify various com-positions, such as fat and muscle. MRI method, one of the HBCA methods, is soaccurate to validate other methods. Applying Whole Body MRI method to largesample improves the reliability of validation, but it also brings di?culties to datagathering and data analysis.Improving MRI technology and utilizing Imaging Processing (IP) technolo-gies can solve these problems. In the past, improving MRI technology was morepopular, but here, I try to proceed with the latter, that is to apply automatic IPtechnologies to MRI method.Chapter 2 and chapter 3 discussed several kinds of IP technologies. In Chapter2, ?rst lists some image preprocessing technologies suitable for MR image. Thenintroduces threshold segmentation, important to MR methods. Then, discussestwo similar methods: Seed Filling Algorithm and Connected Region Labelling.Finally, refers to Canny edge detection technique.Chapter 3 emphases in parameter active contour model: the contour movesunder internal forces and external forces contemporarily, until reach the balance—the contour converge to the true edge. First, reviews traditional snake model andthe dynamical snake model for external forces design. Then introduce the solvingframe—Kass method. Then analysis the classical internal force — Elasticity andBend Force, points out their defects, replace the Elasticity by pressure and dy-namic addition and deletion of nodes.improvements, and delete Bend Force, thenthe Kass method is simpli?ed. Finally, the external force designing: from theclassical image force, distance potential force which enlarge the capture range, toGVF force which enlarge the capture range and able to ?ow into concave edge,the progress of these three potential external forces is remarkable; In addition,the pressure force solved the shortcoming that the contour can only shrink, butthe snake will explosion when shrinks, so replace the pressure with the gradientpotential force of the region bounded by contour, and the result is very good.Chapter 4 sets up MRI method to analysis the body composition. Its keysteps are as follows: Preprocessing Images; Synthesis the seed ?lling algorithmand connected region labelling to extract the interested object; Cut out the re-peated organization to guarantee that the data is uniquely; Using improved activecontour model, moving under GVF force and improved pressure, to extract skele-ton; Threshold segmentation of the fat; Calculate whole body fat volume by ?niteelement method.Finally, in view of image processing, active contour and body compositionanalysis, summarize and look forward to the future separately.