COLOR IMAGE QUANTIZATION USING K-MEANS

Abstract

Image color quantization is a compression technique that aims at reducing the number of colors used to represent an image on a machine. In this work, we will present our application of the K-means algorithm on the color quantization problem. K-means is an unsupervised machine learning algorithm for clustering. The algorithm will form "k" classes (clusters) containing each of them the most homogeneous pixels (with respect to the others belonging to the other clusters) based on the Euclidean distance between them. After loading an image, choosing the number of colors (value of "k"), the tool we have developed in Python, will apply the k-means algorithm and produce another version of the initial image represented only by "k" colors.