【摘要】 将高斯-马尔可夫随机场用于织物纹理的建模.在估计模型参数时,采用不完全 Cholesky共轭梯度法.获得特征参数后,通过纹理合成方法,仿真原始纹理图像.试验表明:仿真织物纹理图像与真实织物纹理图像具有较为一致的视觉效果,从而论证了高斯-马尔可夫随机场模型对织物纹理建模的有效性.更多还原

【Abstract】 Gauss - Markov random field model(GMRF) is introduced to fabric texture analysis. Incomplete Cholesky Conjugate Gradient Method is utilized for estimation of GMRF model parameters which are used as fabric texture features in texture analysis. After the model parameters have been acquired, they are used for synthesis of the crude fabric texture. Experiments show that the resulting simulated fabric texture image shares almost the same visual effects with the corresponding real fabric texture image. The fact proves the algorithms presented in the paper and the rationality of GMRF for texture modeling of fabric.更多还原