细胞图像中一种具有类别选择功能的标识算法

在传统图像处理流程中,在实现目标计数和测量(如细胞计数和细胞参数计算)之前或之后要进行杂质和非同类目标的剔除工作。它一般由形态学运算或人机交互来进行。这里对传统区域标记法加以改造,利用体现细胞主要特征的面积等参数设计分类器,来判定当前目标是属于正确类别还是属于其它,仅对正确类别目标进行标识,避免了形态学运算带来的误差,降低了系统计算量。并给出了算法实现和结果。

An Object Marker Method with Class Selection Function in Cell Image Process

This paper proposes an object marker method with class selection function in t he cell image process.Firstly we use area and other feature to construct a cla ssifier,through which we can judge what category the object belongs to:cell o r other category like speckle;and only mark the correct object,i.e.cell.Com pared with the traditional maker method in image processing,this new method sho rten the MIAS's whole running time and avoid the distortion from morphological process which is before traditional marker method.And the experiment result is given.