论文部分内容阅读
根据平均场神经网络，本文提出了一种二值纹理图象重建算法。纹理图象用马尔可夫随机场模型描述，重建算法是一个自适应过程，递归地进行模型参数估计和图象重建，平均场方法将图象搜索中基于模拟降温的统计松驰方程转化为一组确定性方程，从而有效地提高了计算效率。计算机实验结果证实了重建算法的有效性。
According to average field neural network, this paper presents a binary texture image reconstruction algorithm. The texture image is described by a Markov random field model. The reconstruction algorithm is an adaptive process that recursively estimates the model parameters and reconstructs the image. The averaging field method converts the statistical relaxation equation based on simulated cooling in the image search into a Set of deterministic equations, which effectively improve the computational efficiency. Computer experimental results confirm the effectiveness of the reconstruction algorithm.