摘要
本文提出了一种基于边缘马尔可夫随机场和波尔兹曼机的边缘检测算法,不仅考虑图象的灰度分布信息,而且利用局部的边缘元素结构。边缘检测过程主要包括两个步骤:(1)利用波尔兹曼机学习算法估计模型参数;(2)波尔兹曼机的迭代搜索算法。实验结果证实了算法的可行性。
A new edge detection algorithm based on a Markov random field and a Boltzmann machine is described, where both gray-level information and local edge structures are included. It mainly consists of two steps: (1)Estimation of model parameters using the Boltzmann machine; (2)An iterative search algorithm based on the Boltzmann machine. Experimental results are given to verify the reliability of the proposed method.
出处
《电子学报》
EI
CAS
CSCD
北大核心
1992年第1期74-77,共4页
Acta Electronica Sinica
基金
国家自然科学基金
关键词
图象
分割
边缘检测
参数估计
Image lable, Markov random field, Boltzmann machine, Edge detection, Parameter estimation, Search algorithm
