基于SOFM神经网络与判别分析的肝硬化识别方法的研究

Recognition Method of Liver Cirrhosis Based on SOFM Neural Network and Discriminant Analysis

将SOFM神经网络与医师诊断相结合对腹部MRI进行样本分类,并利用判别 分析方法对肝硬化进行图像识别．应用此方法对随机抽取的40枚腹部MRI图像进行处理,以 7．5％的误判率结果说明了本方法的有效性．

The sample classification of abdomen MRI is carried out by the combining method of SOFM Neural network and doctor＇s diagnosis. Then the liver cirrhosis is recognized based on discriminant analysis. The new technique is applied to recognize 40 abdomen MR images which is randomly abstracted and get the results of 7.5% error rate, which also prove the effectiveness of this new method.