基于纹理特征的关联规则挖掘方法的医学应用

为满足借助医学图像辅助诊断的要求, 提出了一种基于Apriori算法的特征融合算法： 融合图像的纹理特征和医院信息系统(HIS: HospitalInformation System)中病患自然特征。结合剪枝方法建立关联规则库, 实现了一个可以自动将CT(Computer Tomography)图像分为正常与异常两类的原型系统。依据该系统进行了评价实验。实验表明, 通过该算法建立的关联规则库, 对辅助医生诊断具有较好的效果。

Application Research in Medicine Based on Texture Features Association Rules Mining

In order to meet the requirement of medical image auxiliary diagnosis,we present a feature fusion algorithm based on Apriori algorithm: texture features and patient natural features in HIS(Hospital Information System).Accordingly,the combination of pruning methods associated rule base,prototype system for a CT(Computer Tomography) image is divided into normal and abnormal categories.Experiments were evaluated in accordance with t he system,showing that association rules established by the algorithm library,in the auxiliary doctor diagnosed,with good results.