The Study Of Intuitionistic Fuzzy C-Means Clustering Algorithm For Image Segmentation

Fuzzy C-means clustering (FCM) is a typically and frequently used fuzzy clustering algorithm for image segmentation. The principle of the method is simple and the iterative process is capable of adaptive,but this method also has some significant drawbacks:sensitivity to noise and inaccurate to overlapping portion of the image segmentation. In view of this, this article extend fuzzy sets to intuitionistic fuzzy sets and propose an algorithm that intuitionistic fuzzy means clustering considered with spatial information,the intuitionistic fuzzy sets increased non-membership function, that is use the membership and non-membership to describe fuzziness,which is more accurately than the typical fuzziness of fuzzy set.The specific work and mainly contents of this paper are as follows:1. The advantages and disadvantages of the traditional fuzzy C-means algorithm is discussed. Several Fuzzy C-means algorithms based on spatial information is proposed for the shortcoming that FCM algorithm only use of gray information lead to segmentation inaccurate.2. Due to intuitionistic fuzzy set (IFS) is better than fuzzy set (FS) to depict the uncertainty of data, intuitionistic fuzzy sets is introduced into the fuzzy C-means clustering algorithm, we propose an intuitionistic fuzzy C-means algorithm (IFCM). Experiments show that the IFCM algorithm achieved good segmentation results.3. Classify the existing intuitionistic fuzzy entropy formulas and analyzes the advantages and disadvantages of the formulas. For existing problems that entropy formula in dealing with the special circumstances of membership degree is equal to non-membership degree, strictly the law of intuitionistic fuzzy entropy.Propose an improved entropy formula and verify the new entropy is more reasonable to deal with this situation.4. An improved algorithm: spatial information and intuitionistic fuzzy entropy introduced in IFCM algorithm is proposed in order to overcome the shortcoming of traditional FCM algorithm is sensitive to noise. In order to more effectively obtain the optimal solution of the new algorithm we put forward a new algorithm based on SA-PSO and the optimization algorithm is applied to solve the optimal value of the new algorithm.Experiments show that the proposed algorithm for image segmentation achieved good results, effectively avoiding the influence of noise.