基于形状特征变化的凋亡细胞图像识别研究  

Apoptotic cells recognition based on change of shape characteristics

作　　者：高振林[1] 覃玉荣[1] 屈狄[1] 林飞飞[1] 谢朋宇[1] 胡虹梅[1] 

机构地区：[1]广西大学计算机与电子信息学院,南宁530004

出　　处：《计算机应用研究》2011年第3期1186-1187,1190,共3页Application Research of Computers

基　　金：广西科学研究与技术开发计划项目(0719005-1-5);广西科技攻关资助项目(10124001A-41);广西自然科学基金资助项目(0832060)

摘　　要：凋亡细胞识别是癌症等疾病诊断治疗的重要依据,然而凋亡细胞图像识别一直是图像处理领域的瓶颈问题。通过改进的阈值分割和八方向边界描述法提取细胞形状特征,然后基于类圆心点标记和特征参数算法检测其形状变化,较好地解决了凋亡细胞图像识别问题。实验结果表明,该算法适用于形状较复杂的凋亡细胞识别,系统凋亡细胞识别精度和形状识别精度较高,识别时间较短。Apoptotic cell recognition is an important diagnosis and treatment of diseases such as cancer,but it was still a key problem in the image field.First proposed the improved segmentation and 8 directions boundary description strategy to extract the shape characte-ristics of cells.Then detected the changes of the cells＇ shapes by using circle center marking and characteristic parameter algorithm.Solved the problem of recognition of the cells effectively.The results show that this method is appropriate to the apoptotic cells recognition in complicated fields.The apoptotic cells and the shape characteristics could be recognized accurately within a relatively short time in the system.

关 键 词：凋亡细胞 图像识别 形状变化 特征提取 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]