Medical Image Registration Method Research And Its Application To Vocal Cords Image

Medical image registration is a basic work in medical image processing. The data of original multi mage we got always have relative geometric difference when we analyze medical images. The images need to be registered before we analyze them. There are very few studies in domestic and abroad in adapting the medical image registration technology to vocal cords image processing. By this study we can get the data of vocal cords, with which we can judge some people whether they are suitable for artistic vocations relative to vocal cords. It can provide basis for selecting artistic talents at early stage and is also very useful for diagnosing diseases relative to vocal cords. Firstly the paper introduces the usual methods of medical image registration currently and analyzes their strong points and faults. Then on this basis, I present a new and peculiar registration algorithm according to the features and needs of image registration of vocal cords. It divides the registration into two steps: draw the control points and transform images. It draws control points by the combination of selecting points artificially and automatically and divides the procession of selecting points into two steps: rough registration and fine registration. At the former stage it selects control points artificially and selects several points roughly. Then it transforms the registration image into a transitional image with which has roughly the same size by the method of multinomial rectification. At the latter stage it selects control points automatically by the method of grid spreading, then use these high precision control points to decide the parameters of the multinomialtransformation by the method of linear regression, then does geometrical image transt rmation by the method of multinomial transformation to finish the procession of registration. The method is simple and effective. The control points are enough and distributed evenly. The precision of registration is high and satisfies the need of the choosing of artistic vocal talents. At last, the paper presents two methods to improve the registration scheme on the basis of analyzing the experiment result.