K-means: Image compression

[3 P]Apply the k-means algorithm for lossy image compression by means of vector quantization.

• Download the $512 \times 512$ image mandrill.tif (available for download here). Each pixel represents a point in a three dimensional (r,g,b) color space. Each color dimension encodes the corresponding intensity with an 8 bit integer.

• Cluster the pixels in color space using k-means with $k\in\{ 2,4,8,16,32,64,128\}$ clusters and replace the original color values with the indices of the closest cluster centers. You can use the MATLAB function kmeans to do the clustering. Determine the compression factor for each value of $k$ and relate it to the quality of the image. Apply an appropriate quality measure of you choice.

Present your results clearly, legibly and in a well structured manner. Document them in such a way that anybody can reproduce them effortlessly. Send the code of your solution to anand [at] igi.tugraz.at with the subject "MLA SS15 HW11 - <your name>"