Table 6.
Classification performance (%) of the best configurations versus other works using the same dataset for the same task.
| Method | Accuracy | F1-Score | Precision | Recall |
|---|---|---|---|---|
| Aslan et al. [81] CNN | 98.14 | 98.20 | 98.16 | 98.26 |
| Aslan et al. [81] CNN+BiLSTM | 98.70 | 98.76 | 98.77 | 98.76 |
| Chowdhury et al. [54] (NIA) | 97.74 | 96.61 | 96.61 | 96.61 |
| Chowdhury et al. [54] (IA) | 97.94 | 97.94 | 97.95 | 97.94 |
| Maiti et al. [82] (GHE+TBH) | 96.00 | 96.67 | 97.00 | 95.67 |
| Öksüz et al. [83] | 98.30 | 97.61 | 97.43 | 97.78 |
| Progga et al. [84] | n/a | 98.00 | 98.00 | 98.00 |
| Sakib et al. [85] | 96.00 | 97.67 | 98.00 | 97.67 |
| This work-Approach 1 | 98.04 | 98.22 | 98.52 | 97.95 |
| This work-Approach 2 | 97.07 | 97.27 | 98.14 | 96.49 |
NIA: no image augmentation, IA: image augmentation, GHE: global histogram equalisation, TBH: top bottom hat transform.