基于3D卷积神经网络的阿尔茨海默症及轻度认知功能障碍诊断

Analysis of Alzheimer's Disease Medical Images Based on Deep Machine Learning

轻度认知功能障碍(MCI)作为阿尔茨海默病(AD)的预兆,具有相对复杂多变的特点。准确诊断和有效预防MCI具有重要意义。针对AD和正常对照组(CN)、AD和MCI的诊断问题,利用相应核磁共振图像(MRI)影像学数据,提出基于3D卷积神经网络的诊断模型。实验结果表明,AD与CN的分类准确率达96.7%,AUC为0.983,AD与MCI的分类准确率达94.7%,AUC为0.966,该模型在参数量与诊断精度上都具有较高的性能优势。

Mild Cognitive Impairment(MCI),as a precursor of Alzheimer's Disease(AD),is relatively complex and changeable.Accurate diagnosis and effective prevention of MCI are of great significance.For the diagnosis of AD and normal Control Group(CN),AD and MCI,a diagnosis model based on 3D convolution neural network was proposed by using the corresponding MRI imaging data.The experimental results show that the classification accuracy of AD and CN is 96.7%,AUC is 0.983,the classification accuracy of AD and MCI is 94.7%,and AUC is 0.966.The model has high performance advantages in parameter quantity and diagnosis accuracy.