中文摘要：

介绍了一种基于优化水平集的细胞图像分割算法。优化水平集在水平集算法基础上添加了局部熵和灰度变换，以达到突出边缘和去噪的目的。为修正经典OTSU阈值法忽略目标与背景的类内平均距离，创新性地对阈值选择函数进行改进。实验结果表明，相比于传统算法，该算法在正确分割率和运行时间上更优，在复杂的细胞图像分割中具备有效性和可行性。

英文摘要：

A separating algorithm for neuron stem cell images based on level-set segmentation algorithm combined with improved OTSU criterion is presented. Firstly, prove level-set, edge stopping function is constructed based on local entropy and grayscale transformation which can do a good work for overcoming the drawbacks. Secondly, analyze the shortcoming of the OTSU and propose a new threshold function according to the suggestion that make the variance within clusters as a factor of computing the best threshold. Experiments demonstrate that this algorithm has proved high-speed and has a good effect of cell segmention compared to traditional methods.