基于Vision Transformer的阿尔茨海默病分类研究

Research on Classification of Alzheimer’s Disease Based on Vision Transformer

为了有效地提升对阿尔茨海默病(AD)的磁共振成像(MRI)图像分类准确率,提出一种LC(Layer-Cut)-ViT方法。该方法通过引入Vision Transformer(ViT)的自注意力机制对MRI图像进行层切分,使模型能更好地理解图像的全局信息,同时突出切片间的特征关系。此外,通过配准、颅骨分离算法提取MRI图像的脑部组织部分,进一步提升模型的性能。实验结果显示,所提方法对阿尔茨海默病的MRI图像具有较好的分类能力。

To effectively improve the classification accuracy of magnetic resonance imaging(MRI)for Alzheimer’s disease(AD),we propose an LC(Layer-Cut)-ViT method in this paper.This method introduces the self-attention mechanism of the Vision Transformer(ViT)and performs layer-wise segmentation on the MRI images,to enable the model to better understand the global information of the images while emphasizing the inter-slice feature relationships.Additionally,the extraction of brain tissue from the MRI images is further enhanced by employing registration and skull-stripping algorithms,which results in improved performance of the model.Experimental results demonstrate that the proposed method exhibits good classification ability for MRI images of Alzheimer’s disease.