Texture synthesis

Abstract

Texture synthesis is the process of algorithmically constructing a large digital image from a small sample digital image. The fundamental work in this project is to study the Single Resolution Texture Synthesis algorithm, which was developed by Li-Yi Wei and Marc Levoy. To synthesis a new texture, which appears to be generated by the similar stochastic process with the sample texture, the Single Resolution Texture Synthesis algorithm has to use large sampling from the sample texture. However, the larger the sampling procedure is, the lower efficiency will be resulted. Hence, the main task in this project is to develop and implement algorithms which are able to synthesis high visual fidelity textures with faster speed and lower computational cost based on the Single Resolution Texture Synthesis algorithm. Besides, at the final stage of this project, the application of texture synthesis is studied as well.