基于粗糙集理论的肺癌细胞图像识别

The image recognition of lung cancer cells based on rough set theory

肺癌细胞的早期诊断相当困难,肺癌细胞的特征选择依据难以把握.提出根据肺癌细胞的每个特征属性对粗糙集下、上近似集的影响程度作为属性约简的依据,根据约简的结果,再采用扩展的相近关系粗糙集对肺癌细胞进行识别诊断.利用下近似集中的结果进行判断可以提高识别的准确率,利用上近似集中的结果进行判断可以降低肺癌细胞识别的漏诊率.从识别的结果来看,方法行之有效.

Recognizing lung cancer cells at the early stage is a difficult problem in the field of image process and pattern recognition. And feature selection of lung cancer cells is also difficult to handle. Each feature of lung cancer cells was selected by the effect degree to the lower and upper approximation set. And according to the reduce result of feature selection, diagnose of lung cancer cells on the basis of approximation rough set theory was adopted. Lower approximation set of each type of cancer increases the accuracy of recognition, and upper approximation set can decrease the undetected symptom rate. The results of simulation show this method is feasible and effective.