Table 3C.
The comparison of the STARE data set's segmentation results using various segmentation techniques.
| Method | Year | Se | Sp | Acc |
|---|---|---|---|---|
| ECB method (43) | 2012 | 0.7548 | 0.9763 | 0.9543 |
| SP model (19) | 2016 | 0.7867 | 0.9754 | 0.9566 |
| CRF model (20) | 2016 | 0.7680 | 0.9738 | – |
| Cross modality learning (17) | 2016 | 0.7726 | 0.9844 | 0.9628 |
| DSM-UNet (44) | 2018 | 0.7673 | 0.9901 | 0.9712 |
| U-Net+joint losses (23) | 2018 | 0.7581 | 0.9846 | 0.9612 |
| CRF-Net (45) | 2018 | 0.7543 | 0.9814 | 0.9632 |
| SD-UNet (32) | 2019 | 0.7548 | 0.9899 | 0.9725 |
| Three-stage DL Model (25) | 2019 | 0.7735 | 0.9857 | 0.9638 |
| Ipn-v2 and octa-500 (27) | 2019 | 0.7595 | 0.9878 | 0.9641 |
| AA-UNet (34) | 2020 | 0.7598 | 0.9878 | 0.9640 |
| Iternet (36) | 2020 | 0.7715 | 0.9886 | 0.9701 |
| NFN+ Net (46) | 2020 | 0.7963 | 0.9863 | 0.9672 |
| D-GaussianNet (47) | 2021 | 0.7904 | 0.9843 | 0.9837 |
| HDS-Net (30) | 2021 | 0.7946 | 0.9821 | 0.9626 |
| ResDo-UNet (39) | 2021 | 0.7963 | 0.9792 | 0.9567 |
| FPM-Net (proposed) | 2022 | 0.8618 | 0.9819 | 0.9727 |