摘要
把神经网络应用于尿沉渣有形成分的识别,并在神经网络识别的基础上引入模糊推理系统。当遇到神经网络无法正确识别的不规则有形成分时,由模糊推理系统进行推理识别。专家对不规则有形成分的识别经验通过建立多规则、多输入的Mamdani模糊推理模型,融入到识别系统中,使整个识别系统变得智能化。实验结果表明,该方法提高了不规则有形成分的识别率,达到预期的效果。
A new method is proposed to classify the components in urinary sediment images based on the combination of neural network and fuzzy reasoning. If some abnormal components can not be classified by the neural network accurately, fuzzy reasoning is used to classify the components further. Mamdani fuzzy reasoning model which has multi-rules and multi-inputs is created based on the recognition experiences of experts, which makes the recognition system more intelligent. Experimental results indicate that the method can improve the recognizing rates of the abnormal components.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第22期5789-5791,共3页
Computer Engineering and Design
基金
江苏省六大人才高峰基金项目(06-A-027)
江苏省高校自然科学基金项目(06KJD520122)
关键词
尿沉渣
特征提取
神经网络
模糊推理
专家智能
urinary sediment
feature extraction
neural network
fuzzy reasoning
expert intelligent
