Design And Implementation Of Auxiliary Diagnosis System For Alzheimer’s Disease Based On Deep Learning

Alzheimer’s disease is an irreversible neurodegenerative disease that causes tremendous suffering and financial stress to both patients and families.Due to the latent onset and long course of the disease,patients often reach an advanced stage by the time they develop obvious symptoms and cannot be effectively treated.In recent years,deep learning techniques have been widely used in neuroimaging analysis,and the use of deep learning techniques to process neuroimaging for Alzheimer’s disease classification has become an important research topic.Based on this,this paper constructs a deep learning-based Alzheimer’s disease auxiliary diagnosis classification model and implements an Alzheimer’s disease auxiliary diagnosis system to assist Alzheimer’s disease doctors in diagnosis and improve their work efficiency,which has certain practical value.The main work of this paper is as follows:(1)To address the two problems of the current Alzheimer’s disease classification model with less information provided by unimodal data and insufficient feature fusion of multimodal data,this paper proposes a pixellevel feature fusion method based on Cross Attention and constructs an Alzheimer’s disease auxiliary diagnosis model based on it.Firstly,the MPCNN model uses 3D-ResNet1O to extract MRI image features and PET image features respectively in the feature extraction stage.Second,the MPCNN model uses three different fusion methods in the multimodal fusion stage,including feature fusion,decision fusion and pixel-level feature fusion,respectively.Binary classification experiments are conducted on the ADNI public dataset to select the most suitable multimodal fusion method for this paper.Finally,the model is compared with models proposed by other scholars and the model features are output using a visualization algorithm to assist in verifying the validity of the model.(2)Based on the constructed Alzheimer’s disease auxiliary diagnosis model,this paper designs and implements the Alzheimer’s disease auxiliary diagnosis system based on Spring Boot framework in order to meet the actual needs of doctors in the diagnosis process.Firstly,the requirement analysis and outline design of AD-ADS system are introduced in detail.Secondly,the detailed design and implementation of the AD-ADS system includes six modules:personal information management,patient information management,medical history management,scale information management,image management and AD diagnosis management.Finally,a detailed testing of each module of the system is carried out.The main contribution of this paper is to firstly propose a pixel-level feature fusion method based on Cross Attention,on which an MP-CNN model is constructed,and a benchmark model comparison experiment is conducted on the model to achieve a classification accuracy of 93.22%between Alzheimer’s patients and normal controls,which is an improvement in most metrics compared to other classification models.In addition,this paper visualized the output features of the model to prove the effectiveness of the model;then the AD-ADS system was constructed to realize the whole process management of Alzheimer’s patients,which can assist doctors in diagnosing Alzheimer’s disease with high diagnostic accuracy and ease of use,and has a broad application prospect.