基于3DVGGNet轻度认知障碍的分类方法

打开文本图片集

中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2025）23-0048-05

Classification Method for Mild Cognitive Impairment Based on 3DVGGNet

CHEN Guobao1，GONG Xiao² （1.Schoolof Information Engineering， Jiangxi Institute ofTechnology，Nanchang 33oo29，China; 2.Jiangxi Science and Technology Infrastructure Platform Center，Nanchang 33oo03，China）

Abstract：Aimingat the problemof large uncertainty in theartificial feature extractionoperation oftraditional Machine Learning algorithm，aMild CognitiveImpairment （MCI） detection methodbased on3DVGGNetis proposed.Compared with thetraditional Machine Learningalgorithm，this methodhasagreatimprovement inaccuracy，ealizes theautomaticextraction offeatures，andadoptsanend-toend training method.Atthesametime，ithasadeeper networkstructureandabeterceptive feld，andalsohastheadvantageofcontroingthenumberoftrainingparameters.Aimingattheproblemoflongtraining time， thel6-layer network architectureof3DVGGNetisused intheresearch，andthetraining parametersreach 31M.Compared withthetraditionalMachineLearningalgorithm，thetraining timeissavedandthe training speedofthemodelisimproved. Theexperiment is conducted on the ADNI dataset.The results show that the method achieves an accuracy of 85.13% in the classification task ofMCIand NormalControl （NC）groups，providing an effective technical meansfortheearly diagnosis of MCI.

Keywords： cognitive disorder; 3DVGGNet; Deep Learning

0 引言

阿尔茨海默病（AD）的前期表现为轻度认知障碍（MCI），且MCI大概率会进展为阿尔茨海默病[1]。（剩余6159字）