Table 1 Breakdown of the seven-point criteria evaluation dataset29 by management and diagnosis labels and the train-valid-test splits used to train the model.
From: Predicting the clinical management of skin lesions using deep learning
Management labels | Diagnosis labels | Split | Total | ||||||
---|---|---|---|---|---|---|---|---|---|
BCC | NEV | MEL | MISC | SK | Training | Validation | Testing | ||
CLNC | 0 | 268 | 0 | 24 | 4 | 133 | 51 | 112 | 296 |
EXC | 42 | 278 | 252 | 23 | 10 | 235 | 127 | 243 | 605 |
NONE | 0 | 29 | 0 | 50 | 31 | 45 | 25 | 40 | 110 |
Total | 42 | 575 | 252 | 97 | 45 | 413 | 203 | 395 | Â |