Quantum Inspired Automatic Clustering for Multi-level Image Thresholding

Clustering is a simple technique to make partition of given data set into number of clusters. This paper presents an quantum inspired algorithm using GA to automatically find the number of clusters for image data set. The advantage lies in this technique is that no previous information about the data set used for classification is required before hand. The method decides the optimum cluster number on run. The popular evolutionary method called genetic algorithm has been used for generation wise improvement of clustering. CS measure is used as a fitness function in clustering. Effectiveness and accuracy of the proposed technique are demonstrated in terms of standard error found in computation. Finally, desired number of threshold values are heuristically taken from the input image to produce the image after thresholding.