Research On Segmentation Algorithms Of Heterogeneous Medical Image Based On Level Set Method

With the rapid evolution of computer software with hardware and computer graphics theories, Image segmentation has gradually become the focus of attention. In medical research and clinical practice, the diagnosis and treatment of many diseases rely on image segmentation, image segmentation is an important means to help doctors to diagnose accurately. Due to the medical image has the characterizes of complicated structure, noise, weak boundary,and intensity inhomogeneity, the accurate segmentation of the target area is very difficult. Therefore, the research of medical image segmentation algorithm has a great practical significance in the medical field.Image segmentation has become a hot topic. At present, image segmentation based on level set algorithm has been widely concerned by many researchers for its strong local adaptability and high flexibility. In this paper, existing level set segmentation algorithms are detailed described and summarized, and a novel level set method is put forward. The major work of this paper is as follows:1.We illustrated the two types of image segmentation algorithm in medical image, and their advantages and disadvantages are analyzed. We introduced parameter active contour model and geometric active contour model, and focuses on the theory of level set methods and the two kinds of classification of level set method. Finally, we introduced several typical algorithms in the application of level set method in medical image segmentation.2. We propose an improved inhomogeneous medical image segmentation algorithm. Taking into account global and local segmentation capabilities of CV model and LGDF model. We utilized CV model to initialize the contour close to the true boundaries by a preliminary segmentation, then LGDF model is used to attract the contour and stop it at object boundaries. Experimental results show that the accuracy and effectiveness of the method.3. We propose a variation level set method to bias correction and segmentation for images with intensity inhomogeneities. Firstly, we define a localized K-means –type clustering objective function for image intensities in a neighborhood around each point. The cluster centers in this objective function have a multiplicative factor that estimates the bias within the neighborhood. The objective function is then integrated over the entire domain to define the local intensity fitting term. Secondly, in order to drive the evolution of the active contour, a global intensity fitting term is added to the energy function. In the process of the curve evolution, the global intensity fitting force is complementary to each other. Finally, this energy is then incorporated into a variation level set formulation with a level set regularization term that avoids expensive reinitialization of the evolving level set function. Furthermore, we extend this method to three phase level set function for brain MR image segmentation and bias field correction. By using this three-phase level set function to replace the four-phase level set function, we can reduce the number of convolution operations in each iteration and improve the efficiency. In this paper, the algorithm is implemented and the experimental results are given.