基于直方图判别的细胞图像分割

Segmentation for Cellular Images Based on Histogram Differentiation

最优阈值的选取是细胞图像分割中的核心问题,单一的分割方法往往不能适用于各种疏密不一的细胞图像。针对传统的阈值分割方法不能有效地解决稀疏细胞图像分离的问题,该文利用图像直方图的统计特性,提出了一种极小阈值点的搜索方法。并在对直方图进行统计分析的基础上,自动选取恰当的(方法来对不同类型的细胞图像进行分割。实验表明,通过这种方式处理的细胞图像分割效果较好。

The choice for the optimum threshold is the core of the segmentation for cellular images. A single method is hardly suitable for all images with various kinds of density. Because the usual methods is difficult to make a proper segmentation for the sparse cell's vision, this article utilizes image statistics characteristic of histogram, puts forward one light search method that threshold value order extremely. At the same time on the basis of the thing that carry on statistical analysis to the histogram, choose the suitable method of cutting apart automatically according to different kinds of cellular images. Experimental results show that one can obtain better segmentation of image by using this approach.