| Image segmentation plays an important role in image analysis and image understanding, which has a direct influence to the consequent effects. In recent years, active contour model has become one of the most popular methods on image segmentation, which transform image segmentation to a procedure of seeking energy functional minimization.Based on the research of active contour model, this dissertation analyzes the characteristics and applications of each model, combines active contour model with new ideas and methods, and then research new segmentation methods. The main work is as the following:(1) The important role of image segmentation and current methods are described, focused on the principle and processes of active contour model. Due to the different parameters in the implementation, the mathematical basis of active contour model-the level set are compared and discussed.(2) Aiming at the shortcomings of GAC model, a grey relation based image segmentation method is proposed, which replace the traditional image gradient information with gray correlation function to build a new edge function of GAC model. Compared with the traditional gradient information, gray relational degree is able to express fuzzy edges and lead a fast convergence. Experimental results show that this method is superior to the traditional GAC and LBF model based segmentation methods in terms of segmentation effect.(3) After analyzing the basic principles and characteristics of the basic CV model, a grey relation based image segmentation method is proposed, which regards gray relational degree as edge information and further introduces it to C-V model. This method employ CV model to adjust the global information, while gray relational degree improves the segmentation effect in case of blurred edges as a kind of edge information. Experimental results show that this method is superior to the traditional GAC and basic C-V model based segmentation methods in terms of segmentation effect. |
PDF Full Text Request