Table 1 Proportion of ocular surface slit-lamp images covering the cornea area in the training, validation, and test datasets.
From: Deep learning for identifying corneal diseases from ocular surface slit-lamp photographs
| Â | Retrospective dataset n (%) | Prospective dataset n (%) |
|---|---|---|
Normal Subjects | 870 (16.3%) | 87 (17.1%) |
Cataract | 1,860 (34.9%) | 160 (31.4%) |
Infectious keratitis | 845 (15.9%) | 86 (16.9%) |
Non-infectious keratitis | 785 (14.7%) | 81 (15.9%) |
Corneal dystrophy or degeneration | 550 (10.3%) | 54 (10.6%) |
Corneal Neoplasm | 415 (7.8%) | 42 (8.2%) |
Total | 5,325 | 510 |
