Year | Study | Dataset | Method | Layers | Error | Time (S) |
---|---|---|---|---|---|---|
2010 | [48] | Healthy adults | GT + DP | 8 boundaries | Layer thickness difference MAE GCL-IPL = 0.77 \(\pm\) 0.65 px OPL = 1.48 \(\pm\) 1.05 px (px = 3.29 um) | 9.74 |
2013 | [50] | Healthy | GT + energy minimization | 6 boundaries | MAE IPL-INL = 4.67 \(\pm\) 0.83 um SE IPL-INL =  − 3.59 \(\pm\) 0.93 um (1px = 3.9 um) | 18 |
2015 | [49] | Healthy and DME | KR + GT + DP | 8 boundaries  + fluid | Layer thickness differences GCL-IPL = 4.84 \(\pm\) 5.12 um OPL = 6.35 \(\pm\) 6.11 um (px = 3.87 um) | 11.4 |
2018 | [51] | Healthy children’s and AMD | RNN-GS | 7 boundaries of healthy, 3 of AMD | SE INL/IPL = − 0.13 \(\pm\) 1.10 px OPL/INL = − 0.10 \(\pm\) 1.31 px MAE INL/IPL = 0.56 \(\pm\) 0.95 px OPL/INL = 0.69 \(\pm\) 1.12 px (px = 3.9 μm) | 145 |
2021 | [52] | Chiu dataset, Healthy, AMD, CSCR and DME | GT + WGD | 8 boundaries | SE IPL-INL =  − 8.12 \(\pm\) 5.59 INL-OPL =  − 8.33 \(\pm\) 7.68 OPL-ONL =  − 1.41 \(\pm\) 3.51 MAE IPL-INL = 12.67 \(\pm\) 3.49 INL-OPL = 7.88 \(\pm\) 6.96 OPL-ONL = 1.84 \(\pm\) 3.28 | – |