针对卵胞浆内单精子显微注射过程中人类细胞质、注射针和卵细胞极体的分割问题，首先根据分割目标形状的差异性，采用骨架形态学算子提取感兴趣区域;接着采用改进的模糊聚类方法，依据局部灰度和局部方差特征对边界模糊、噪声较大的预处理图像进行粗分类;然后在模糊聚类的基础上设定水平集初始化位置和演化控制参数，实现细胞质、极体和注射针的分割;最后针对细胞质内的部分注射针图像，采用直接拟合标记出完整注射针图像． 实验测试结果表明，文中算法可以实现人类卵细胞的细胞质、极体和注射针的正确分割．

Aiming at the segmentation of human cytoplasm，injection pipette and egg polar body in the process of intra-cytoplasmic sperm injection (ICSI)，firstly，the operator of mathematical morphology is adopted to detect the region of interest according to the difference in shapes of segmentation targets． Secondly，according to local gray level and local variance，an improved fuzzy clustering method is used to implement the rough classification of pre-processing images with fuzzy boundaries and noises． Then，the level set algorithm is used to carry out the segmen-tation of cytoplasm，polar body and injection pipette，with its initialization location and controlling parameters be-ing set according to fuzzy clustering results． Finally，line fitting is employed to mark the whole injection pipette from the local images of injection pipette inserted in cytoplasm． Experimental results show that the proposed algo-rithm can implement the segmentation of human oocyte cytoplasm，polar body and injection pipette correctly．