中文摘要：

目的利用细胞图像特殊的结构特点，设计出一种新的基于其梯度信息和灰度信息的主动轮廓模型细胞图像分割算法。方法首先将原细胞图像和该图像的梯度图像以一定比例相加得到新的图像，然后，设置第一个水平集函数的初始值，进行水平集函数的迭代求得细胞质和背景的轮廓线。最后用相同的方法求得细胞质和细胞核的轮廓线。结果与现有的水平集方法得到的分割结果比较，本文所提算法主要优点是能够更加精确地分割出细胞质、细胞核和背景3部分。结论细胞独特结构使得主动轮廓模型分割算法结合梯度信息能够更好的实现细胞图像分割。

英文摘要：

Objective To design a new active contour mouel cell image segmemauon algorlmm Dasect on the gradient information and gray information according to the special structural features of cell image. Methods The original cell image and the gradient image were added at a certain proportion to obtain a new image, and the initial value of the first level set function was set for the new image. Then the level set function was iterated to achieve the contour line between cytoplasm and background. Finally, the same method was used to achieve the contour line between the cytoplasm and nucleus. Results As compared with the existed level set methods, the new algorithm could obtain more accurate segmentation results of the cytoplasm, nucleus and background. Conclusion The special structure of cell makes active contour model algorithm combined with gradient information segment cell image more accurate.