Multiple studies presented satisfactory performances for the treatment of various ocular diseases. To date, there has been no study that describes a multiclass model, medically accurate, and trained on large diverse dataset. No study has addressed a class imbalance problem in one giant dataset originating from multiple large diverse eye fundus image collections. To ensure a real-life clinical environment and mitigate the problem of biased medical image data, 22 publicly available datasets were merged. To secure medical validity only Diabetic Retinopathy (DR), Age-Related Macular Degeneration (AMD) and Glaucoma (GL) were included. The state-of-the-art models ConvNext, RegNet and ResNet were utilized. In the resulting dataset, there were 86,415 normal, 3787 GL, 632 AMD and 34,379 DR fundus images. ConvNextTiny achieved the best results in terms of recognizing most of the examined eye diseases with the most metrics. The overall accuracy was 80.46 ± 1.48. Specific accuracy values were: 80.01 ± 1.10 for normal eye fundus, 97.20 ± 0.66 for GL, 98.14 ± 0.31 for AMD, 80.66 ± 1.27 for DR. A suitable screening model for the most prevalent retinal diseases in ageing societies was designed. The model was developed on a diverse, combined large dataset which made the obtained results less biased and more generalizable.

Fundus images play a fundamental role in the early screening of eye diseases. On the other hand, as deep learning provides an accurate classification of medical images, it is natural to apply such techniques for fundus images. There are many developments in deep learning for such image data but are often burdened with the same common mistakes. Training data are biased, not diverse and hidden to the public. Algorithms classify diseases, which suitability for screening could be questioned. Therefore, in our research, we consolidate most of the available public data of fundus images (pathological and non-pathological) taking into consideration only image data relevant to the most distressing retinal diseases. Next, we apply some well-known state-of-the-art deep learning models for the classification of the consolidated image data addressing class imbalance problem occurring in the dataset and clinical usage. In a conclusion, we present our classification results for diabetic retinopathy, glaucoma, and age- related macular degeneration disease, which are urgent problem of ageing populations in developed countries.