An Improved Intuitionistic Fuzzy C-Mean Segmentation Algorithm Based on Intuition-Point Method
The traditional Fuzzy C-Means(FCM)algorithm does not fully utilize the constraints of the algorithm when solving the membership function and cluster center expression,resulting in the obtained membership function not converging to the optimal value,reducing the dependence between target pixels and neighboring pixels and cluster centers,and affecting the effectiveness of image segmentation.Aiming at this problem,the penalty function was de-signed by using the idea of the inner point method,the penalty function was introduced into the objective function of the FCM algorithm,and then the KKT condition was used to find the membership under the full constraint condition.The fuzzy set was then generalized to the intuitive fuzzy set to introduce non-membership and hesitation,further opti-mizing the membership matrix and refining the uncertain information in the image.The segmentation experiments show that the improved algorithm not only enhances noise immunity,but also protects the details of the image.Conclusion:Compared with the FCM algorithm,the segmentation accuracy of this algorithm is improved by at least 3.11%,and the division coefficient is increased by 9.29%.Compared with the FCM_S2 algorithm,the segmentation accuracy of the proposed algorithm is increased by about 0.5%,and the division coefficient is increased by 2.14%.
Image segmentationInterior point methodPenalty functionIntuitionistic fuzzy set
NETL
NSTL
万方数据
