Application Of Image Segmentation Based On Improved FCM Algorithm

Image segmentation is an important process of image analysis and recognition,which plays essential role in the succeeding image processing.Accurate,fast and effective image segmentation can make image identification more precise and high-efficiency.Therefore,improving the effect and efficiency of entire image segmentation can greatly reduce the cost of image processing.Fuzzy C-means clustering algorithm(FCM),as a general method,is often applied in current image segmentation field.However,it exists the problem such as easily falling in local extremum,being sensitive to noise and mis-segmenting images.In this paper,in order to solve the locality of classical fuzzy C-means clustering algorithm,we propose two improved fuzzy C-means clustering algorithms for image segmentation.First,we present a GA-based fuzzy C-means clustering algorithm to optimize the problem of easily falling in local optimum.Then we employ fuzzy C-means clustering algorithm combined with random walk to enhance its noise-resistant ability and segmentation effect.Our simulation is conducted on UCI dataset and multi-class image data,by comparing with classical FCM algorithm,which shows the proposed two algorithms have significant advantages in segmentation effect and accuracy.This paper first introduces the purpose and significance of this research,and then classifies the commonly used image segmentation methods,and describes the current research situation of each category at home and abroad.Secondly,an improved genetic algorithm(RGA)is introduced.Based on this,a fuzzy C-means clustering(RGA-FCM)algorithm based on improved genetic algorithm is proposed to solve the problem that the classical FCM algorithm is easy to fall into the local optimal and the UCI data set Experiments on the upper and multiple sets of image data to verify the effectiveness of the RGA-FCM algorithm.Finally,a random walk algorithm is proposed to solve the problem that the classical FCM algorithm only considers the gray level information in the image segmentation and proposes the FCM(spatial fuzzy c-means,SFCM)algorithm which integrates the spatial information and experiments on different image data The superiority of the proposed algorithm.