Texture synthesis by genetic algorithm

Abstract

One way to synthesis texture in a fast and easy way is image quilting proposed by Efros and Freeman in 2001. This research brings the adaptive search power of genetic algorithm and combines it with the concept of image quilting to propose new texture synthesis algorithm. The proposed GA is ran on many different images from standard texture sets. Visual comparison of our proposed GA with image quilting algorithm is considered. The texture results generated by the proposed GA are roughly comparable in quality to those generated from Efros and Freeman algorithm.