基于K均值聚类算法的宫颈癌细胞分割方法  

Segmentation for Cervical Cancer Cells Based on K-Means Clustering Algorithm

作　　者：赵英红[1] 洪雅玲 孙存杰[1,2] 

机构地区：[1]徐州医学院医学影像学院,江苏徐州221004 [2]徐州医学院附属医院,江苏徐州221006

出　　处：《临床医学工程》2014年第9期1089-1090,共2页Clinical Medicine & Engineering

摘　　要：目的探究一种对宫颈癌细胞显微图像进行彩色分割的方法,辅助临床诊断。方法采用彩色聚类分割方法,并结合K-means聚类算法实现细胞图像的有效分割。结果实现了宫颈癌细胞图像的有效分割,色彩信息得到最大保留。结论该实验结果有利于辅助专家进行宫颈癌病理诊断,并为后期研究宫颈癌细胞特征识别奠定基础。Objective To explore an effective segmentation algorithm method for microscopic image of cervical cancer cell, to assist clinical diagnosis. Methods Effectively segmentation for cell image was obtained by color clustering algorithm, combining improve, d K-means clustering algorithm. Results Cervical cancer cell image was segmented effectively, and kept the color information as much as possible. Conclusions The result is helpful for pathological diagnosis and characteristics identification for later research on cervical cancer ceils.

关 键 词：宫颈癌细胞 K—means聚类 彩色图像分割 

分 类 号：R318.6[医药卫生—生物医学工程]