Table 4 Number of validation patients that the per-patient model can not classify since the prediction confidence is too small for that specific value of the threshold t
From: Endoscopy-based IBD identification by a quantized deep learning pipeline
Threshold (t) | Unpredicted |
|---|---|
0.85 | 0 |
0.87 | 0 |
0.88 | 0 |
0.90 | 0 |
0.91 | 0 |
0.93 | 0 |
0.94 | 1 |
0.96 | 1 |
0.97 | 2 |
0.99 | 4 |