复杂背景下的细胞图像分割技术研究  

Cell Pattern Recognition Technology Researchunder Complicated Background

作　　者：陈宇[1] 范影乐[1] 庞全[1] 

机构地区：[1]杭州电子科技大学自动化分院,杭州310018

出　　处：《计算机工程与应用》2005年第7期42-43,59,共3页Computer Engineering and Applications

基　　金：国家自然科学基金(编号:60374047;60302027);浙江省自然科学基金项目(编号:ZD0205;602127);浙江省教育厅科研项目(编号:20030620)资助

摘　　要：针对复杂背景下的细胞图像分割问题,考虑到传统阈值分割方法的局限性,提出了一种自适应局部阈值分割算法。该算法的基本思想是将首次分割得到的目标区域按照一定准则进行划分,分为已经正确提取的目标细胞和包含有多个细胞及部分背景的模糊区域,然后将各个模糊区域作为新的图像分别进行再次分割,从中提取出目标细胞。实验结果表明,该算法能够稳定地实现目标分割提取,具有较好的鲁棒性。In view of the limitations of traditional simple threshold algorithm under the complicated background,an algorithm for adaptive local segmentation is proposed.The main technique of the algorithm consists in:the object areas resulting from first segmentation are divided into correct goal cell areas;fuzzy areas include a lot of cells and some background according to certain criterion;then each fuzzy area regarded as new image goes on for the second segmentation respectively,the goal cell is drawn out from it.The experimental result indicates that the algorithm can realize steadily that the goal is cut apart and drawn,and is stupid and excellent better.

关 键 词：图像分割 复杂背景 局部阈值 自适应 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]