Texture Synthesis Algorithm Research And Application

Texture synthesis has been one of the hottest topics in the areas of photorealistic rendering and non-photorealistic rendering, and been widely used in the fields of virtual reality, computer vision etc. Texture is used to represent the rich details on the surface of the object. It can be also used to depict the natural phenomenon that has repeated features. Texture synthesis technology in these areas has a broad prospect. It becomes a hotspot in the areas of graphics and image processing.First, this paper introduces the status of the study and relative basic concept in texture synthesis. Next, it takes the Markov Random Field as the foundation to summary and analysis all the algorithms in the past of this article, following this, it present an improved algorithm based patch, and applies the algorithm to 3D-curved surface texture mapping. This paper contains the following work: (1) we introduce and summary some classic algorithms about texture synthesis through different category, we also carry on the elaboration to the question the areas of Chinese Painting simulation and image inpainting. (2) We analysis some algorithms in sample-based texture synthesis and multi-sampled based texture synthesis, and have a study of Texture transferring texture mixing and constrained multi-based texture synthesis. This paper present an improved concrete algorithm based patch. (3) It summaries some algorithms in texture mapping about cylinder and spherical surface, We apply the algorithm in (2) to 3D-curved surface texture mapping, and have some calculation results.