KMeans Segmentation - MEX

Version (5.7 KB) by Ahmad
Given N data elements of R dimensions (N x R matrix), it segregates the n elements into k clusters
Updated 23 Jun 2010

View License

KMEANSK - mex implementation (compile by mex kmeansK.cpp
Also an equivalent MATLAB implementation is present in zip file

Performs K-means clustering given a list of feature vectors and k. The argument k indicates the number of clusters you want the data to be divided into. data_vecs (N*R) is the set of R dimensional feature vectors for N data points. Each row in data_vecs gives the R dimensional vector for a single data point. Each column in data_vecs refers to values for a particular feature vector for all the N data points. The output data_idxs is a N*1 vector of integers telling which cluster number a particular data point belongs to. It also outputs centroids which is a k*R matrix, where each rows gives the vector for the cluster center. If we want to segment a color image i into 5 clusters using spacial and color information, we can use this function as follows:

% r = i(:,:,1);
% g = i(:,:,2);
% b = i(:,:,3);
% [c r] = meshgrid(1:size(i,1), 1:size(i,2));
% data_vecs = [r(:) g(:) b(:) r(:) c(:)];
% [ data_idxs centroids ] = kmeansK( data_vecs, k );
% d = reshape(data_idxs, size(i,1), size(i,2));
% imagesc(d);

Cite As

Ahmad (2024). KMeans Segmentation - MEX (, MATLAB Central File Exchange. Retrieved .

MATLAB Release Compatibility
Created with R2009b
Compatible with any release
Platform Compatibility
Windows macOS Linux

Inspired: Sparsified K-Means

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!
Version Published Release Notes

Comments update. Plus shifted some functionality to a utils file

Included some input checking!