摘要
提出了一种基于结构性信息的判断纹理块匹配误差的新方法:颜色—结构差异算法。该方法同时考虑了纹理块之间颜色误差和结构误差对纹理块相似性的影响,弥补了单纯考虑颜色误差的不足,保证了图像间结构信息的连贯性,从而更好地保证了合成后的纹理在视觉上的连贯性。
A new difference measure for texture synthesis, named Intensity-Structure Difference, is presented in this paper. This measure considers the intensity difference and structure difference simultaneously when computing the similarity within patches. It successfully overcomes the drawbacks the Intensity Difference method has. The structure continuity within a texture is ensured, and the perceptual coherence in synthesized texture is retained.
出处
《计算机应用研究》
CSCD
北大核心
2006年第4期175-176,224,共3页
Application Research of Computers
关键词
马可夫随机域
纹理合成
颜色-结构差异
Markov Random Field
Texture Synthesis
Intensity-Structure Difference(ISD)