Image Processing Based On Level Set Method

With the development of multi-media technology, images have been an important way for human to express themselves or communicate with others. Different images express different information. So how to deal with and to analysis the information which the image expresses become more and more important. As two parts of the technology of image processing, image segmentation and image denoising have become the hot topics. The technology of image processing has been used in many fields, such as the space exploration, remote sensing, the geology exploration, biomedical engineering, communication engineering and so on. In every field of them mentioned above, we will have so many pictures to process, to recognize and to analyze.Medical images are so important to the real life. The processing and analysis of medical images determine the diagnosis of the patients. Segmenting the images in an accurate way can help the doctors make or change a correct treatment plan, and improve the accuracy of medical diagnosis.First of all, a new model based on the theory of curve evolution has been proposed in this paper. It is improved based on the RSF Model. We increase a weight term about the area. And we construct a function about the information of the gradient of image grayscale instead of a constant. In this way, the adaption of the model has been stronger than the older one. The curve on the image can run independently and purposely. In this paper, the samples used in the experiments are clinical images. Then we do the experiment in the matlab, and what the results shows us is better than the old model and which proves to be a more efficient method of image segmentation.Secondly, we combine an image segmentation model and an image denoising model to deal with the images with noise. What we have done above is to reduce the affect that the noise produces, and to segment the image correctly and precisely. The image segmentation model used here is the one that we mentioned before. Also we do the experiment in the matlab, and what the results show us is better than those we did before. The images which have been processed become smoother, and can be segmented more precisely.