A few-shot learning-based eye diseases screening method

OBJECTIVE: This study aims to construct a brand-new ophthalmic disease screening task and establish a practically valu-able ophthalmic disease screening model in the case of insufficient data. MATERIALS AND METHODS: The main methods are as follows: firstly, we mixed da-ta from different sources (these data may come from different cameras, including different fun-dus diseases) to get a new dataset. Based on this dataset, we conducted subsequent experi-ments on fundus multi-disease screening. How-ever, in the past public datasets, each dataset often only corresponded to the screening diag-nosis of one disease. Secondly, we proposed a method to simulate the characteristics of differ-ent fundus cameras by using a method based on style transfer, and to augment the training da-ta, so that the model could learn the features of ophthalmic diseases in a more comprehensive way. Finally, a robust disease screening mod-el based on few-shot learning was construct-ed on the combined dataset, and compared with benchmark algorithms. RESULTS: We focused on the study of eye disease screening methods based on the met-ric-based few-shot learning model, data aug-mentation methods, and focus on key tech-nologies such as data augmentation based on style transfer. Experiments have shown that our method can significantly improve the general-ization ability of the disease screening model. CONCLUSIONS: By introducing few-shot learning theory and data augmentation based on style transfer into ophthalmic disease screening, the generalization ability of the model is greatly improved, and it has certain practical value.