Research of Texture Synthesis Algorithm Based-On Patch Sampling
|School||Henan University of Science and Technology|
|Course||Applied Computer Technology|
|Keywords||texture synthesis MRF model based-on patch sampling Gauss pyramid spiral path|
Texture synthesis from Sample is a new texture synthesis technology which is developed after texture mapping and procedural texture synthesis method. It is used to solve the questions of joint, distortion, and parameter adjustment which appear in the traditional method. It shows a widespread application prospect in the image editing, the damage image packing, the data compression, the network data transmission, the large-scale scene production as well as third dimension image and non-third dimension image and so on, and it is one of the domain research hot spots in the computer graphic, the computer vision and the image processing and so on.This article first reviews several typical two-dimensional texture synthesis algorithms which are based on sample images, discusses the basic thoughts of these algorithms and the insufficiency existed in these algorithms. Based on these algorithms, we propose the corresponding improvement algorithms, this article main work manifests in:Firstly, we apply the thought of multi-resolution texture synthesis to the patch-based sampling algorithm. The multi-resolution has the effect of“neighborhood expands”, that is, using small match boundary then we can obtain similar synthesis effect with the single resolution synthesis algorithm which uses big boundary, meanwhile employ KD tree to search the best match block in sample image. This method can reduce computational complexity for patch-based sampling algorithm.