Table 13.

FIVE-CLASS DR GRADING ON THE EYEPACS BY OUR MODELS COMPARED WITH OTHER DL MODELS (UNIT: %).

Models	Acc	QW-Kappa
CF-DRNet [66]	83.10	-
Wang et al. [58]	85.70	83.70
DeepMT-DR [59]	86.90	85.20
CABNet [16]	86.68	86.07
Li et al. [67]	-	83.46
CoT-XNet [61]	84.10	-
FA-Net	86.60	84.58
FA+KC-Net+R1	86.63	84.66
FA+KC-Net+R2	86.78	84.93