from
K Means Algorithm with the application to image compression
by Jason Rebello
This program uses the K means clustering algorithm to group the pixels in an image
|
| computeCost(X, idx, centroids, K) |
function cost = computeCost(X, idx, centroids, K)
% computeCost is used to compute the final cost of all the points belonging
% to their respective clusters
% Initialize variables
cost = 0;
% Compute cost
for i=1:K
temp = find(idx==i);
Xtemp = X(temp,:); % Get all points belonging to respective cluster
cost = cost + (1/length(Xtemp))*sum((sum(((bsxfun(@minus,centroids(i,:),Xtemp)).^2),2))...
.^(-1/2));
end
% Formula : {sqrt[(q1-p1)^2 + (q2-p2)^2 +...]}/{number of points belonging
% to cluster}
% q = q1 + q2 + ...
|
|
Contact us