From: Identification of ocular refraction based on deep learning algorithm as a novel retinoscopy method
 | Training set | Validation set | Test set |
---|---|---|---|
No. of patients | 2769 | 791 | 394 |
Sex, (M/F) | 1548/1221 | 351/440 | 173/221 |
Age (y), mean (SD) | 18.35 (6.50) | 18.72 (7.34) | 18.94 (7.22) |
RM (No. of images) | 5511 | 1575 | 787 |
CM (No. of images) | 5511 | 1575 | 787 |
A-CMa (No. of images) | 1420 | 406 | 202 |
Sphere, mean (SD) | − 3.77 (2.04) | − 3.95 (2.05) | − 3.95 (2.09) |
Cylinder, mean (SD) | − 0.82 (0.61) | − 0.81 (0.60) | − 0.83 (0.63) |
Axis (W/A/O) | 2920/1543/1048 | 882/504/189 | 519/204/64 |
SE, mean (SD) | − 4.18 (2.11) | − 4.36 (2.12) | − 4.17 (2.12) |
High myopia | 36.9% | 39.6% | 36.2% |
Moderate myopia | 28.7% | 29.1% | 28.7% |
Mild myopia | 34.4% | 31.3% | 35.1% |
Intraocular pressure (mmHg) | 16.1 (2.01) | 15.9 (2.16) | 16.4 (1.78) |
Uncorrected distance visual acuity (LogMAR) | 0.68 (0.25) | 0.69 (0.21) | 0.69 (0.22) |
Centre corneal thickness | 551.57 (30.93) | 555.17 (22.18) | 547. 28 (22.61) |
K1 | 42.41 (1.25) | 42.35 (1.33) | 42.41 (1.31) |
K2 | 43.75 (1.41) | 43.99 (1.36) | 43.96 (1.43) |