采用DenseNet模型的AD自动分类方法

Automatic classification of Alzheimer's disease using DenseNet model

为研究深度学习算法对阿尔茨海默病分类的准确性,提出密集卷积神经网络方法,对阿尔茨海默病进行分类.利用预处理后的数据训练密集卷积神经网络结构,并分类阿尔茨海默病和认知正常者.测试结果表明,文中方法获得的分类准确率为98.91%,分类阿尔茨海默病和轻度认知障碍的准确率为94.54%,准确率较其他算法有一定提升,为阿尔茨海默病的精准分类提供了一种有效的解决方案.

In order to study the accuracy of deep learning algorithms in classifying Alzheimer′s disease,a dense convolutional neural network(DenseNet)method was proposed.The preprocessed data is used to train a dense convolutional neural network structure and classify Alzheimer′s disease and people with normal cogni-tion.The test results show that the classification accuracy obtained by this method is 98.91%.The accuracy rate of classifying Alzheimer′s disease and mild cognitive impairment is 94.54%,which is significantly im-proved over other algorithms and provides an effective solution for the accurate classification of Alzheimer′s disease.