一种基于Mamdami模糊逻辑推理的血细胞分类方法

摘要

Based on manual classification of blood cells images, a model of automatic classification of blood cells is constructed. With the application of Mamdami fuzzy logical inference, this model helps the completion of the blood cells' membership of the cell attribute sets inferring from cell image attribute space. Then, the membership of the blood cells is mapped to image semantic space. At last, the blood cells are automatically classified. Nonlinear transformation of the semantic mapping from attribute space to image semantic space is critical in this model. This automatic classification method of the blood cells is realized and testified by the initial experiments.%通过模拟血细胞图像人工识别过程,建立自动识别分类模型,该模型利用Mamdami的模糊逻辑推理来完成从细胞图像特征空间到细胞属性特征的隶属度,再从属性隶属度到语义空间的鉴别映射,最后在语义空间中实现血细胞的分类.语义鉴别映射是实现从细胞图像特征空间到细胞语义空间的非线性变换,是细胞识别方法的关键.本文提出的血细胞自动分类识别方法是基于血细胞认识原型分类方法的具体实现.通过初步实验表明该方法是可行的.