Research on the application of vision Transformer model in 14-visual-field classification of fundus images

As an important basis for clinical diagnosis of ophthalmology, fundus images play a key role in early detection, assessment and diagnosis of various ophthalmic diseases. However, the traditional fundus image analysis relies on the professional experience of ophthalmologists, and the process is time-consuming and subjective, which is difficult to meet the needs of modern medical treatment. Based on the Vision Transformer (ViT) model, this study proposes an innovative 14-visual-field classification method for fundus images, aiming to break through the limitations of traditional methods and improve the efficiency and accuracy of fundus image analysis. By constructing a dataset covering rich cases and combining the global perception and efficient data utilization of ViT model, the automatic classification of fundus images is realized, which provides a new technical path for the intelligent diagnosis of eye diseases. This method can help promote the early screening and treatment of eye diseases, so as to better protect people’s visual health.