Research Article

Multiscale U-Net with Spatial Positional Attention for Retinal Vessel Segmentation

Table 4

The results of different algorithms in the DRIVE dataset.

DatasetMethodsYearSe (%)Ac (%)AUC (%)F1-score (%)

DRIVER2U-Net [30]201877.9295.5697.8481.71
U-Net [30]201875.3795.3197.5581.42
LadderNet [31] 201878.5695.6197.9382.02
DUNet [32]201978.9496.9798.56N/A
DEU-Net [33]201979.4095.6797.7282.70
AG-Net [34]201981.0096.9298.56N/A
IterNet [12]201977.3595.7398.1682.05
BCDU-Net [35]201980.0795.6097.8982.24
Tang et al. [36]202081.6095.5497.99N/A
Lü et al. [37]202080.6295.4797.39N/A
SA-UNet [38]202082.1296.9898.6482.63
Zhang et al. [13]202081.5196.9598.63N/A
RVSeg-Net [39]202081.0796.8198.17N/A
Proposed method202183.1696.9998.7682.91