Table 2.
Confusion matrices for deep convolutional ensemble and board-certified ophthalmologists showing the mean (and SD) per cent agreement between the predicted labels against the ground-truth labels over the test set
| Deep convolutional ensemble | Ophthalmologists | ||||||||
| Normal | DR | Glaucoma | AMD | Normal | DR | Glaucoma | AMD | ||
| Ground-truth labels | Normal | 23.8% (0.8%) |
1.2% (0.8%) |
0.0% (0.0%) |
0.0% (0.0%) |
21.9% (2.3%) |
1.3% (1.4%) |
1.4% (2.1%) |
0.4% (0.5%) |
| DR | 6.6% (1.1%) |
18.2% (1.1%) |
0% (0.0%) |
0.2% (0.4%) |
9.7% (5.1%) |
12.4% (4.5%) |
1.0% (1.4%) |
1.9% (1.1%) |
|
| Glaucoma | 6.6% (2.3%) |
0.2% (0.4%) |
18.2% (2.7%) |
0.0% (0.0%) |
5.6% (4.8%) |
0.9% (1.6%) |
17.1% (3.8%) |
1.4% (0.5%) |
|
| AMD | 3.2% (0.4%) |
2.8% (1.5%) |
0.0% (0.0%) |
19% (1.2%) |
0.7% (0.8%) |
2.6% (1.5%) |
0.4% (0.5%) |
21.3% (1.4%) |
|
Green cells indicate agreement between the ground-truth labels and predictions by the deep convolutional ensemble or ophthalmologists, and red cells similarly indicate disagreement
AMD, age-related macular degeneration; DR, diabetic retinopathy.