Table 2 Proposed model class wise result in percentage.
From: A hybrid XAI-driven deep learning framework for robust GI tract disease diagnosis
Class | Sample size | Precision | Recall | F1-score | FNR |
|---|---|---|---|---|---|
Dyed-lifted-polyps | 500 | 0.90 | 0.95 | 0.93 | 0.05 |
Dyed-resection-margins | 500 | 0.97 | 0.90 | 0.93 | 0.10 |
Esophagitis | 500 | 0.88 | 0.96 | 0.92 | 0.04 |
Normal-cecum | 500 | 0.94 | 0.98 | 0.96 | 0.02 |
Normal-pitylorus | 500 | 0.97 | 0.97 | 0.97 | 0.03 |
Normal-z-line | 500 | 0.94 | 0.84 | 0.89 | 0.16 |
Polyps | 500 | 0.95 | 0.90 | 0.93 | 0.10 |
Ulcerative-colitis | 500 | 0.95 | 0.96 | 0.95 | 0.04 |
