SERR-U-Net: Squeeze-and-Excitation Residual and Recurrent Block-Based U-Net for Automatic Vessel Segmentation in Retinal Image

Table 2

The division of datasets for training, validation, and test.

Database	Training	Validation	Test
DRIVE	20 images (190000 patches)	4 images (38000 patches)	20 images
STARE	10 images (95000 patches)	2 images (19000 patches)	10 images