摘要
提出一种基于K近邻法的模糊决策的肝纤维化CT图像分类方法。对提取的图像频域特征向量用模糊加权K近邻法进行分类 ,其中引入隶属度函数对由于各种噪声和扫描参数的变化引起的特征值的不确定性进行描述。本研究结果表明模糊技术的应用提高了分类器的识别率和鲁棒性。
The paper presents a classifier of K-nearest-neighbor based on fuzzy decision.The classifier uses a memˉbership function to describe the uncertainty of features extracted from the spectrums of images aroused by various noises and the changes of scanning parameters.In the classifier the values of fuzzy function are used as weights.The results demonstrated that the technology of fuzzy improves rates of detection and the robust of classifier with CT images of hepatic fibrosis.
出处
《医疗设备信息》
2004年第11期15-16,39,共3页
Information of Medical Equipment
关键词
模糊技术
K近邻法
特征提取
隶属度函数
technology of fuzzy
K-nearest-neighbor
feature extraction
membership function
