Purpose of the guideline: |
In recent years, AI algorithms have shown good performance in the analysis and recognition of lesion signs in the macular area of fundus color images. As time goes by, screening systems of fundus lesion signs suitable for primary screening will definitely appear in the future. This guideline is intended to provide reference opinions for the definition of the lesion signs in macular area, AI application scenarios, algorithm model building, the establishment of a dataset and labeling, screening system architecture and function design, and image data acquisition. Meanwhile, it can standardize the application of AI system in the primary screening of lesion signs in macular region, and promote the overall improvement of the early screening level of fundus disease in China. |