摘要： 为减少混凝土CT图像分割过程中信念信息传递的计算复杂度，提出一种基于多尺度信念传播的图像快速分割算法。根据标准信念传播算法的特点，构造金字塔状栅格图进行消息传输，从而加快消息收敛速度、减少迭代次数，同时利用双向图技术减少信念信息的数量。实验结果表明，该算法能提高图像分割率和分割精度。

关键词: CT图像, 图像分割, 多尺度, 信念传播, 细观结构

Abstract: In process of concrete CT image segmentations, the compute complexity is higher when belief messages are transmitted. Aiming at this problem, this paper proposes a fast segmentation algorithm based on multiscale belief propagation. According to the feature of belief propagation algorithm, it structures grid figure with pyramid shape for message transmission, so as to speed up the convergence speed and reduce iteration times. It uses ambigram technology to reduce faith the quantity of information. Experimental results show that the proposed method can improve efficiency and precision of the image segmentation.

Key words: CT image, image segmentation, multiscale, belief propagation, meso-structure

中图分类号: 

• TP391