Research On Texture Synthesis Based On Sample Texture

Texture synthesis technology is one of the hottest research topic in the fields of currently computer graphics, computer vision and image processing, which has widely apply value. Since TSFS has not only overcome weaknesses of traditional texture mapping and also avoided of the tedious process caused by physical model which is used to synthesis texture, it is gaining more and more attention in the research area. Several classical algorithms on which we applies some advanced method on are introduced here in this paper, including a synthetical algorithm on texture synthesizing based on Texture Detail intensity calculating by using co-occurrence matrix analysis is proposed in the paper. Classifying the texture into the random texture part and the structural texture part using the co-occurrence matrix analysis to synthesis the two different parts respectively.To do this, we combine the advantages of two approaches so that we can speed up the time and gain a better result. Meanwhile, we presents an efficient algorithm applies the multiresolution method into the random texture synthesizing. Further, this paper provides experimental results to demonstrate the influence caused by the different neighborhood size applied to the experiment.