Color quantization with nature inspired computing

In this thesis, two different nature inspired optimization algorithms are applied to the Color Quantization (CQ) problem. CQ is a technique that is used to decrease the number of distinct colors in an image. In many applications, an image with a limited number of colors provides more effective implementations either in software or hardware. However, this process reduces the visual quality of the image. Therefore, one of the major concerns in CQ is keeping the distortion of the quantized image as low as possible. CQ is considered as an NP-Hard optimization problem.In the first part of this study, for the first time, Intelligent Water Drops (IWD) algorithm is adopted to Basic CQ. The proposed algorithm IWD-CQ was tested on a set of standard test images. The comparison results against eight approaches revealed that IWD-CQ algorithm achieves promising results in terms of the visual quality of quantized images. The algorithms used for comparison include five widely used conventional methods; Uniform, Popularity, Median-Cut, Octree and K-Means and three prominent Artificial Intelligence methods; Self Organizing Maps, Ant Colony Optimization and Simulated Annealing. The results show that none of the mentioned algorithms achieve lower quantization errors than the IWD-CQ algorithm.Second part of the study, investigates a Perceptual CQ approach which is based on a model of Human Vision System (HVS). The model aims to simulate the property of human?s eye in perceiving an average color of the adjacent pixels. The related study was carried out using Differential Evolution (DE) algorithm as the optimizer. Moreover, tuning of the DE parameters for the proposed method is covered. Outcomes of the proposed algorithm, DE-PCQ, are discussed and compared with results of previously mentioned five conventional methods applying the same perceptual approach. Comparison results demonstrate distinct superiority of DE-PCQ in terms of the perceptual quality of quantized images.