Research Of Image Segmentation Algorithm Based On KFCM And Improved CV Model

The Image segmentation is an important tasks in the computer science andartificial intelligence field, which is very challenging and one of the indispensable linkin image processing courses.Recently the research based on the algorithm of imagesegmentation has been unabated, on one hand fruitful results has been gained. Manyalgorithms have been improved respectively on one or a few aspects such as correct rate,computational efficiency and quality, robustness, excessive reliance of imagesegmentation.It is proposed in this paper based on KFCM and improve CV model Split Bregmanimage segmentation method to solve initial value, computational efficiency androbustness of image segmentation algorithm. A series of experiments verify thefeasibility and stability of the algorithm, and the obvious improvement has gained inthese problems. The research work controls the following aspects:（1）Make a summary of existed segmentation algorithms. According to thecurrent domestic and foreign existing image segmentation algorithms which are fromimportant references and a lot of related books, a more comprehensive summary hasbeen made.（2）The related image segmentation algorithms are expressed based on fuzzyc-means clustering. The previous work is described on improvement and developmentof the FCM clustering algorithm,when improved clustering algorithm isto achieve the achievements and shortcomings of image segmentation. finally thedetailed analysis of KFCM algorithm used for image preprocessing stage is done.（3） Improved algorithm of image CV model segmentation.As CV modelhas the properties of the curve,which adapts totopology changing,it be used in imagesegmentation, discussing its advantages and disadvantages. it can be improved on therecognition rate and computational efficiency.（4）Split Bregman algorithm of image segmentation based KFCM algorithm andimproved CV model. It is expounded the basic principle and process of Split Bregmanalgorithm, while it gives its steps to build process based on KFCM algorithm andimproved CV model.Some experimental and existing algorithm results are comparedand analyzed by a series of simulation experiments the next wok to do is that the relatedalgorithm will be perfected. （5）Based on the research of image segmentation algorithms, it is analyzed andsummarized. Sum up experience and insufficient to indicate the direction of futureresearch, discussing the direction of improved segmentation algorithms.The article is an experimental environment in the MATLAB2010A,experimentswith the image database.Many simulation experiments verify its effectiveness andstability of the algorithm.