Research Image Segmentation Method Based On C-V Model

In image processing and analysis Image segmentation plays an important role, it’s Image processing and analysis’s premise. It has a great significance, which also has a great challenge. As one of the major branches of image segmentation, medical image imaging technology is also Rapidly evolving. In modern medicine, Medical image information processing increase an important role, and The medical image segmentation is a key technology of computer aided diagnosis system of medical. Medical image has features of diversity and complexity, traditional image segmentation algorithm is not ideal, in many cases, we can’t get the result we want. So many scholars research it.Image segmentation by theory of partial differential equations is a new technology only recently thirty years. It bases on a lot of theoretical foundations of mathematics With the characteristics other algorithms don’t have.Active contour model is a main algorithm of image segmentation by theory of partial differential equations, it incorporates a variety of information of the image, the image edge segmented by it is a closed curve. During the same period, the level set methods have been proposed. These two methods Combine one new method, which is called geometric active contour models, with their advantages. This study is to research C-V model algorithm, one kind of geometric active contour models.CV model is a classic geometric active contour model based on the area information, it can process image without significant edges well. However, for some noise image, it can’t handle, the results is good or not based on the initial position of the contour curve, what’s worse, in it’s calculation process, it needs to do periodic initialization constantly to cause it costs much time.In response to these shortcomings, this article propose two improved algorithms, one is based on the Otsu method, it can deal with various types of image more effectively, with better noise immunity and robustness, and it takes less time and reduces the requirements for initial contour position. The other is to add constraint items, in addition to the process of periodic initialization of contour curves, it can save a lot of time, and reduces the requirements for initial contour position.