Research On Texture Synthesis Algorithm And Its Applications

Texture synthesis from Sample (TSFS) is one of the hottest areas in Computer Graphics, Computer Vision and Image Processing research. Given a sample texture, texture synthesis seeks to synthesize a new texture, which is perceptually similar to the given sample and can ensure that its output has sufficient variations. Base on previous research efforts on TSFS, this dissertation mainly present some new ideas of TSFS, which are helpful for improving the quality of synthesized results and accelerating synthesis speed.Firstly, we review the research results of TSFS and introduce its basic model, principle and typical algorithms. Secondly, after studying the technology of texture synthesis over manifold surfaces deeply, we extend the 2D Wang Tiles to three-dimensional space and present a real-time texture synthesis algorithm over arbitrary surfaces using Triangle Tiles. Because triangle tiles are pre-built, there is no dynamic calculation, which makes our method efficient enough and works in real-time. In the last part of this dissertation, the Global Texture Optimization algorithm based on Expectation Maximum method is well studied, then we revise the original global texture optimization method and propose a novel algorithm of image inpainting. Experiment results show that our method can get graceful results both in image inpaiting and removing redundant object in images.