Table 13 Comparative training, validation, and test accuracy & loss—DenseNet121 (Kaggle Binary Class & NDB-UFES Multiclass OSCC Dataset).
From: Deep visual detection system for oral squamous cell carcinoma
Metric | Kaggle Dataset (Epochs = 20, Batch Size = 64) | NDB-UFES Dataset (Epochs = 20, Batch Size = 64) |
|---|---|---|
Train loss | 3.0689 | 1.2197 |
Training accuracy | 0.8914 | 0.5447 |
Validation loss | 2.9701 | 1.2192 |
Validation accuracy | 0.8651 | 0.5478 |
Test loss | 2.9852 | 1.2174 |
Test accuracy | 0.8690 | 0.5476 |
