Table 4 Training and testing set for lymph node models.
From: Accurate recognition of colorectal cancer with semi-supervised deep learning on pathological images
Model | Class | Dataset-Pcam | ||
|---|---|---|---|---|
Training set | Testing set | |||
Labeled | Unlabeled | |||
Pcam-1%-SSL | Tumor | 1311 | 129,761 | 16,384 |
Non-tumor | 1311 | 129,761 | 16,384 | |
Total | 2622/1%a | 259,522/99%b | 32,768 | |
Pcam-5%-SSL | Tumor | 6554 | 124,518 | 16,384 |
Non-tumor | 6554 | 124,518 | 16,384 | |
Total | 13,108/5%c | 249,036/95%d | 32,768 | |
Pcam-1%-SL | Tumor | 1311 | – | 16,384 |
Non-tumor | 1311 | – | 16,384 | |
Total | 2622/1% | – | 32,768 | |
Pcam-5%-SL | Tumor | 6554 | – | 16,384 |
Non-tumor | 6554 | – | 16,384 | |
Total | 13,108/5% | – | 32,768 | |
Pcam-100%-SL | Tumor | 131,072 | – | 16,384 |
Non-tumor | 131,072 | – | 16,384 | |
Total | 262,144/100%e | 32,768 |
