Table 1 Number of ultra-widefield fundus images in the training, validation, and test datasets.

From: Deep learning for detecting retinal detachment and discerning macular status using ultra-widefield fundus images

 

Training set no. (%)

Validation set no. (%)

Test set no. (%)

Rhegmatogenous RD

1169 (16.0)

244 (15.9)

256 (16.3)

Tractional RD

50 (0.7)

15 (1.0)

11 (0.7)

Exudative RD

54 (0.7)

12 (0.8)

7 (0.4)

Recurrent RD after scleral buckling

59 (0.8)

14 (0.9)

16 (1.0)

Recurrent RD after vitrectomy

72 (1.0)

16 (1.0)

14 (0.9)

RDs

1404 (19.2)

301 (19.3)

304 (19.3)

Non-RD

5919 (80.8)

1255 (80.7)

1268 (80.7)

Total images of RD/non-RD

7323 (100)

1556 (100)

1572 (100)

Macula-on RD

360 (29.1)

77 (28.9)

80 (30.1)

Macula-off RD

879 (70.9)

189 (71.1)

186 (69.9)

Total images of macula-on/off RD

1239 (100)

266 (100)

266 (100)

  1. RD retinal detachment