【摘要】 目的建立肝癌CT影像人工神经网络辅助诊断模型。方法对所收集的肝癌(110例)和其它肝占位疾病(123例)的CT影像进行预处理,提取各病例CT的影像学特征并以此作为网络训练样本集和网络评价样本集。建立肝癌CT影像误差逆传播(back propagation,BP)神经网络辅助诊断模型,用训练样本集对网络进行训练后,将评价样本集通过网络仿真,用仿真输出数据计算该神经网络用于肝癌CT影像诊断的灵敏度、特异度和准确度。结果网络的训练过程显示,第12步时训练停止,样本集划分合理,对网络的训练效果良好。根据网络的仿真输出数据,计算出神经网络用于肝癌CT影像诊断的灵敏度和特异度分别为98%和96%,诊断准确度为97%。结论该人工神经网络可以应用于肝癌CT影像的临床诊断。更多还原

【Abstract】 Objective To establish CT accessory diagnostic model based on artificial neural network (ANN) in hepatocellular carcinoma(HCC).Methods CT images of HCC (110 cases) and other liver occupied disease (123 cases) were preprocessed.CT features of each CT image were extracted that actecd as the training and evaluating database.CT accessory diagnostic model of ANN for diagnosis of HCC was constructed with a back-propagation(BP).After trained the network with traning database,the evaluating database was through the network.The sensitivity,specificity and accuracy of the network in diagnosing HCC according to the output data.Results The network training process showed that the training lasted 12 epoch.The classification of database was proper.Effect of training on network was very well.The sensitivity,specificity and accuracy of ANN were 98%,96% and 97% respectively for diagnosis of CT diagnosis of HCC.Conclusion The ANN can be used for clinic diagnosis of HCC on CT.更多还原