基于骨架的血细胞图像分离算法

A New Blood Cell Image Split Algorithm Based on the Skeleton

文章从细胞图像骨架的角度出发,研究和发展了一种新的粘连细胞分离算法。其主要思路是:首先采用细化算法提取细胞图像的骨架,并计算骨架各点的边界距离值,然后利用骨架边界距离函数波谷的特性将相互粘结的细胞分离。新算法提出了一种计算骨架边界距离值的新方法用于分析骨架波谷的特性,并在理论和实践的基础上,研究和发展了细胞图像中粘结细胞的分离准则。为证明该算法的有效性,选取了上百种不同的细胞图像进行了测试。实验结果表明,新算法能有效、快速地对细胞图像中相互粘结的细胞进行合理分割。

This paper presents a new algorithm for blood touched cell image segmentation based on the skeleton.The main procedures are:obtaining skeletons of the cells regions on an image,using distance function to analyze the properties of the skeletons and to split touched cells.In the algorithm,a new method to get skeleton distance for analyzing skeleton properties is proposed and studied,and splitting rulers for touched cell are developed based on theoretically analysis and practice experience.To check the efficiency of the algorithm,hundreds different types of images are tested.The testing results show that the algorithm splits touched cells reasonably and quickly.