Table 21 Comparative classification report of oral squamous cell carcinoma (OSCC)—ResNet50 (Kaggle vs. NDB-UFES Dataset).
From: Deep visual detection system for oral squamous cell carcinoma
Dataset | Epochs | Batch Size | Matric | Precision | Recall | F1-Score | Support |
|---|---|---|---|---|---|---|---|
Kaggle | 10 | 64 | Normal | 0.66 | 0.40 | 0.50 | 389 |
OSCC | 0.57 | 0.79 | 0.66 | 390 | |||
Accuracy | 0.60 | 779 | |||||
Macro Avg | 0.61 | 0.60 | 0.58 | 779 | |||
Weighted Avg | 0.61 | 0.60 | 0.58 | 779 | |||
NDB-UFES | 20 | 64 | OSCC | 0.57 | 0.55 | 0.56 | 167 |
With dysplasia | 0.55 | 0.76 | 0.64 | 301 | |||
Without dysplasia | 0.00 | 0.00 | 0.00 | 97 | |||
Accuracy | 0.56 | 565 | |||||
Macro Avg | 0.37 | 0.44 | 0.40 | 565 | |||
Weighted Avg | 0.45 | 0.56 | 0.50 | 565 |
