Level Set Method Of Image Segmentation Based On Improved Region Models

Image segmentation is a crucial step in the field of computer vision.In the field of scientific research,it has been widely studied by scholars at home and abroad and has formed a variety of algorithms.Among them,the image segmentation based on level set method has become the hotspot and develop rapidly in recent years due to its unique algorithm advantages such as free topology transformation,strong mathematics foundation,multi-information communicability,and good expandability.In this paper,we studied the region-based level set method,and improved the existing deficiencies in the existing models.Finally,we combined many kinds of artificial synthetic images and real images in engineering practice to verify the feasibility of the model.The innovation works made in this paper are as follows:Aiming at the phenomenon that the existing local models are easy to fall into local minimum and causes segmentation failure when segmenting images with intensity inhomogeneity,an improved model VLIF which can effectively solve this problem is proposed.This model incorporates the idea of maximumvariance between classes and adds a class-to-class variance energy term based on the LIF model.By maximizing the difference between the target and the background in the neighborhood of all points on the evolution curve,the pseudo edge points which lead the evolution curve into a local minimum are eliminated.Then the evolution curve been driven to stay at the correct target boundary.In order to prove the validity of the proposed model,a series of experiments were conducted in this paper.The results show that the improved model can effectively solve the problem caused by the local minimum in local models and improves the accuracy of segmenting result in intensity inhomogeneity and complex images.A weight-self adjustment active contour model combined with the image global information and local information is proposed.The local entropy of the image is used to establish an accurate and reliable index for measuring image intensity information.In order to achieve fully automated segmentation,we can obtain the intensity distribution in the process of image segmentation to guide the matching of different functional energy items in the model adaptively in the real time.Compared to the traditional fixed-parameter linear adjustment model based on trial and error,this paper selects a more quantified and reasonable way to achieve the dynamic nonlinear adjustment of global items and local items and realizes fast and accurate segmentation of various types of images.Finally,through a series of experiments,it is shown that the proposed model can achievefast and accurate segmentation for different types of intensity inhomogeneous and noisy images with high stability and insensitive to the position of the initial contour.