Table 4.
Comparison between the results of the BFCN and other methods on the DRIVE dataset.
| Method | Year | Sensitivity | Specificity | Accuracy | F-measure |
|---|---|---|---|---|---|
| Singh and Srivastava [8] | 2016 | 0.7594 | 0.9723 | 0.947 | — |
| Zhang et al. [10] | 2015 | 0.7812 | 0.9668 | 0.9504 | — |
| Lázár and Hajdu [38] | 2015 | 0.7646 | 0.9723 | 0.9458 | — |
| Zhao et al. [39] | 2017 | 0.782 | 0.979 | 0.957 | — |
| Zhang et al. [40] | 2017 | 0.7861 | 0.9712 | 0.9527 | 0.7953 |
| Yan et al. [41] | 2018 | 0.8282 | 0.9738 | 0.9609 | — |
| Alom [42] | 2018 | 0.7792 | 0.9813 | 0.9556 | 0.8171 |
| Zhuang [43] | 2018 | 0.7856 | 0.9810 | 0.9560 | 0.8202 |
| Wang et al. [44] | 2019 | 0.7940 | 0.9816 | 0.9567 | 0.8270 |
| Samuel and Veeramalai [21] | 2019 | 0.8282 | 0.9738 | 0.9609 | — |
| Ours | 2019 | 0.8124 | 0.9822 | 0.9627 | 0.8294 |