|国家科技期刊平台
首页|期刊导航|现代信息科技|基于迁移学习的脑部磁共振图像的阿兹海默病分类的应用研究

基于迁移学习的脑部磁共振图像的阿兹海默病分类的应用研究OA

Application Research on Alzheimer's Disease Classification of Brain Magnetic Resonance Image Based on Transfer Learning

中文摘要英文摘要

随着社会人口老龄化问题的加剧,阿兹海默病,越来越影响人们的生活质量和家庭幸福,也造成了巨大的社会负担.利用人工智能技术对AD进行早期诊断有助于预防或减缓AD病程,减轻家庭和社会负担.现有文献表明,基于磁共振图像的AI分类算法可用于AD早期诊断.针对基于磁共振图像的AD分类问题,设计实现了两种分类迁移学习方法,分别是微调方法和时域视觉提示方法,并通过在公开数据集上验证,证实这些方法的分类精度得到提升.

As the aging of the social population intensifies,Alzheimer's disease(AD),is increasingly affecting people's life quality and family happiness,and has also caused a huge social burden.Using Artificial Intelligence technology to conduct the early diagnosis of AD can help prevent or slow down the course of AD and reduce the burden on families and society.Existing literature shows that AI classification algorithms based on magnetic resonance images can be used for early diagnosis of AD.Aiming at the AD classification problem based on magnetic resonance images,this paper designs and implements two classification Transfer Learning methods,namely fine-tuning method and time-domain visual prompting method.Through verification on public data sets,it is confirmed that the classification accuracy of these methods is improved.

吕姝瑶

北京航空航天大学,北京 100191

计算机与自动化

阿兹海默病磁共振图像医学图像分类迁移学习

ADmagnetic resonance imagemedical image classificationTransfer Learning

《现代信息科技》 2024 (016)

39-43,48 / 6

10.19850/j.cnki.2096-4706.2024.16.009

评论