kernel_distance_kld.m

高斯滤波器 matlab toolbox For GMMs and Gaussian kernels

function K = kernel_distance_KLD(g1, g2, N)

s = kernel_samples(g1, N);
w1 = kernel_evaluate(g1, s);
w2 = kernel_evaluate(g2, s);

%K = sum(log(w1) - log(w2)) / N;

ii = find(w1 ~= 0 & w2 ~= 0);
K = sum(log(w1(ii)) - log(w2(ii))) / length(ii);