Study Of Image Segmentation Technology Based On Fuzzy Clustering

An image is a kind of significant information that humanity acquires.Putting the image into regions with characteristics and extracting useful target play an important part in our daily life and scientific researches.However,it is difficult to extract the interest target of an image,for the image is always influenced by many factors in the process of creation,transmission and record.Therefore, scientific researches on image segmentation become more and more imperative.Because of complexity and uncertainty of image information,it is difficult to know the image objects and their locations in the processing of image segmentation.Fuzzy cluster analysis with the ability to describe the problem in conjunction with fuzzy clustering methods used in image segmentation can get a better effect in comparison with the traditional methods.Fuzzy C-means(FCM) clustering algorithm is a classical one in fuzzy cluster analysis algorithms, using iterative optimization of the objective function to obtain the fuzzy partition data sets,and is good at convergence.FCM algorithm can avoid setting the threshold,solve a number of branches of the partition problem,and suit images with uncertainty and ambiguity.As an unsupervised clustering algorithm FCM algorithm does not need manual intervention and is very significant for the automatic of image segmentation.Therefore,the application of FCM algorithm to image segmentation,being a hot subject in image processing,has a certain practical value.In this paper,focusing on improved fuzzy clustering algorithm and its application in image segmentation,the following work is done.First,because the traditional fuzzy c-means algorithm does not consider the space information of pixels and is sensitive to noise,an improved algorithm for image segmentation based on fuzzy c-means clustering is proposed.The spatial distance between a pixel and the cluster center is calculated by the membership matrix of the neighboring pixels,and a new distance is determined by the spatial distance and the Euclidean distance.This new distance feature and the improved algorithm based on fuzzy c-means clustering are used in image segmentation.The experimental results of two types of noisy images show that the proposed algorithm is effective to get the target image and more robust to the noises,and has fast convergence.Secondly,because the image segmentation algorithm is complexity to calculate the feature set, lacks enough robustness to noise and outliers,and costs much time in computation,a new algorithm for image segmentation based on fast fuzzy c-means clustering is proposed in this paper.In order to reduce the number of iterations,the algorithm chooses the peak value of gray-level histogram for the image as the prior cluster center.To enhance the noise immunity of image segmentation,the clustering of centre pixel is influenced by the neighbor average values and median values.The algorithm reduces the time of each iteration step by the statistical histogram of the image.The experimental results show that the proposed algorithm is more effective to constrain the noise and faster than many other segmentation algorithms.Finally,the regularity of change of misclassified pixels with the parameterαis analyzed when the image is corrupted by Gaussian noise,salt and pepper noise,mixed noise,respectively.The number of iterations is reduced due to the selecting of the initial centroid and the clustering based on the gray histogram of the image.Usually,after several iterations convergence and stability will be achieved.The number of misclassified pixels by the proposed algorithm is rapidly decreasing while the value ofαis increasing and relative smoothness appears afterα＞4.