基于Markov因果邻域系统的纹理综合算法研究  

Algorithm study on texture synthesis based on Markov causal neighborhood

作　　者：李莉[1] 曹继华[1] 

机构地区：[1]天津职业技术师范学院电子系,天津300222

出　　处：《天津职业技术师范学院学报》2004年第2期29-31,共3页Journal of Tianjin Vocational Technical Teachers'college

基　　金：天津自然科学基金项目(023600111).

摘　　要：讨论了一种基于Markov因果邻域系统的纹理综合算法,它使用Markov/Gibbs模型描述纹理,避免了复杂概率函数,通过确定的搜索过程逐像素产生纹理,减小了运算量,提高了纹理综合效率,使纹理综合可以应用在受限纹理综合和运动纹理综合等领域。This paper discusses an algorithm based on Markov causal neighborhood, which not only uses Markov/Gibbs model but also generates the new texture image pixel by pixel through a deterministic searching process. This avoids the computational demand of probability sampling so as to develop the efficiency of texture synthesis. As a result, it texture synthesis constrained texture synthesis and motion synthesis possible.

关 键 词：纹理综合 因果邻域 受限纹理综合 运动纹理综合 

分 类 号：O211.62[理学—概率论与数理统计]