Multi-scale interactive network with artery/vein discriminator for retinal vessel classification
For DRIVE-AV， Se, Sp, and Acc are 98.99%, 93.15%, and 97.79%, respectively
The 100 fundus images dataset has been released.
Password：00di
The labeled data has been normalized. Users can handle it by themselves according to the usage situation.
JingFeiHu GuangWu HuaWang