fcm.rar_FCM分割_fcm_fcm图像分割算法_fuzzy segmentation_聚类

function [C, dist, J] = fcm(X, k, b,m,n) %C 中心 %X n行p列数据集 % J 目标函数 %k 聚类中心数 % dist 隶属矩阵 %b 混合程度 iter = 0; [N, p] = size(X); P = randn(N, k); P = P./(sum(P, 2)*ones(1, k)); J_prev = inf; J = []; while true %mtimes iter = iter + 1; t = P.^b; C = (X'*t)'./(sum(t)'*ones(1, p)); dist = sum(X.*X, 2)*ones(1, k) + (sum(C.*C, 2)*ones(1, N))'-2*X*C'; t2 = (1./dist).^(1/(b-1)); P = t2./(sum(t2, 2)*ones(1, k)); J_cur = sum(sum((P.^b).*dist))/N; J = [J J_cur]; if norm(J_cur-J_prev, 'fro') < 1e-3 %用不同的范数计算 A和A‘的积的对角线和的平方根，即sqrt(sum(diag(A'*A))) break; end if iter>100 break; end fprintf('#iteration: %03d, objective function: %f\n', iter, J_cur); J_prev = J_cur; end