摘要
图像边缘的建模与检测是图像分割、目标区域的识别等图像分析领域中重要的基础。本文采用统计建模的方法对图像进行边缘建模、检测,对基于“线场”与不使用线场的“特殊集簇形式”的两种图像建模及检测方法进行总结和比较,并且指出第一种方法使用的“线场模型”可以转化为第二种方法的“特殊集簇形式求点元的相应势能”模型,并对各自的实验效果进行了分析。
Edge modeling and detection are important in image segmentation and object region recognition in image anlysis fields. This paper presents statistical methods for image edge modeling and detection using line process and special cliques with no adequate line elements, besides transforms the effect of the dual MRF of line elements into effective potentials in the energy function of a Gibbs distribution for pixels, and compars the results of the models.
出处
《计算机与现代化》
2007年第3期77-80,96,共5页
Computer and Modernization
