Alzheimer's disease is an irreversible neurodegenerative disease of the brain.Deep learning technology is used to as-sist doctors is of great significance to improve the efficiency of identifying patients with Alzheimer's disease.Combining the three-di-mensional convolution technology,densely connected convolutional networks(Densenet)and a simple,parameter-free attention module for convolutional neural networks,referred to as the attention module of SimAM,this paper designs a 3D-SAMDensenet al-gorithm to classify the magnetic resonance images(MRI)of the brain.When using this model to classify Alzheimer's disease(AD)and normal people(NC),mild cognitive impairment(MCI)and normal people(NC),better classification results are obtained com-pared with the traditional and deep learning classification methods,and greater advantages are obtained than AD/NC classification in MCI/NC classification with small image differences.
维普
万方数据