Peatland Fire and Smoke Detection on Malaysian Peatland Fire and Smoke Dataset (test)
91.6AccuracyWHT-ResNet50
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| WHT-ResNet50Transfer Learning=Yes, No. of Parameters=20,580,290, Training iterations=252026.03 | 91.6 | 92.9 | 90.17 | 91.5 | |
| SwinTransformerTransfer Learning=Yes, No. of Parameters=27,520,892, Training iterations=252026.03 | 91.1 | 91.9 | 89.7 | 90.8 | |
| ResNet50Transfer Learning=Yes, No. of Parameters=23,512,146, Training iterations=252026.03 | 88.9 | 90 | 86.5 | 88.2 | |
| EfficientNet-B5Transfer Learning=Yes, No. of Parameters=28,361,274, Training iterations=252026.03 | 88.6 | 89.4 | 87 | 88.2 | |
| HTMA-ResNet50Transfer Learning=Yes, No. of Parameters=11,797,826, Training iterations=252026.03 | 87.9 | 88.1 | 87.5 | 87.8 | |
| WHT-ResNet50Transfer Learning=No, No. of Parameters=20,580,290, Training iterations=252026.03 | 77.2 | 78.2 | 76 | 77.1 | |
| SwinTransformerTransfer Learning=No, No. of Parameters=27,520,892, Training iterations=252026.03 | 72.2 | 71.9 | 71.4 | 71.9 | |
| ResNet50Transfer Learning=No, No. of Parameters=23,512,146, Training iterations=252026.03 | 71.4 | 70.9 | 69.2 | 70.1 | |
| EfficientNet-B5Transfer Learning=No, No. of Parameters=28,361,274, Training iterations=252026.03 | 71.2 | 69.8 | 70.4 | 70.4 | |
| HTMA-ResNet50Transfer Learning=No, No. of Parameters=11,797,826, Training iterations=252026.03 | 69.6 | 69 | 66.4 | 67.6 |