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

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.