Table 6.
Past work on the diagnosis of human burns using various ML and DL.
| Study | Model | Classification | Dataset Size | Accuracy (%) |
|---|---|---|---|---|
| Abubakar et al.15 | Pre-trained ResNet50 | Binary Class | 2080 images | 95.43 |
| Pre-trained VGG16 | 85.63 | |||
| Smith et al.16 | Pre-trained ResNet50 | Binary Class | 1360 images (Caucasians dataset) | 99.3 |
| 540 images (African dataset) | 97.1 | |||
| Ugail et al.44 | ResNet101 | Binary Class | 1360 images | 95.9 |
| Buhar et al.17 | VGG-Face | Binary Class | 1420 images | 95.208 |
| Yadav et al.40 | SVM | Binary Class | 94 segmented images | 82.43 |
| Shin et al.45 | SSL | Multiclass | 170 images | 70.0 |
| Rahman et al.46 | SVM | Multiclass | 500 images | 93.3 |