A Reinforcement-Learning Approach to Color Quantization

• DOI: 10.6180/jase.2011.14.2.07
• Publication Date: 2011-06-01
• Authors: Chien-Hsing Chou; Mu‐Chun Su; Yuxiang Zhao; Fu‐Hau Hsu
• Keywords: Color quantization; RGB color model; Color balance; Color depth; Color histogram; High color; Linde–Buzo–Gray algorithm

Color quantization is a process of sampling three-dimensional color space (e.g. RGB) to reduce the number colors in image. By reducing discrete subset known as codebook or palette, each pixel original image mapped an entry according these palette colors. In this paper, reinforcement-learning approach proposed. Fuzzy rules, which are used select appropriate parameters for adaptive clustering algorithm applied quantization, built through reinforcement learning. comparing new method with on 30 images, our shows improvement 3.3% 5.8% peak signal noise ratio (PSNR) values average and results savings about 10% computation time. Moreover, we demonstrate that learning efficacious well efficient way provide solution problem where there lack knowledge regarding input-output relationship.