Table 2.
Breakdown of data used for the convolutional neural network (CNN)
| Score Feature |
Negative | Positive | Ratio of negative to positive |
|---|---|---|---|
| CC | None, Impossible to score | Segmental, Global | |
| 8893 | 1209 | 7.36 | |
| IMM | (−), Impossible to score | (+) | |
| 2782 | 7320 | 0.38 | |
| MLysis | (−), Impossible to score | (+) | |
| 9990 | 112 | 89.20 | |
| F-Cre | (−) | (+) | |
| 8611 | 1491 | 5.78 | |
| IVVP | (−), Impossible to score | (+) | |
| 9958 | 144 | 69.15 |
Each number shows the number of images contained in the score. CC, capillary collapse; F-Cre, fibrous crescent; IMM, increased mesangial matrix; IVVP, increased vasculature around the vascular pole; MLysis, mesangiolysis.