Automatic Detection of AMD and DME Retinal Pathologies Using Deep Learning

Table 5

Obtained results on local dataset.

Model	Accuracy	Precision	Recall	score
Fine-tuning of Xception	96.77%	96.66%	97%	97%
Features extraction from “sepconv2_bn”+CNN	98%	94.66%	94.66%	94.66%
BCNN (Xception, Xception)	98.56%	96.33%	96.66%	95%
Fine-tuning de Inception-resnet-v2	96.11%	95%	95%	95%
Features extraction from “block8_1_ac”+CNN	97.12%	98.33%	98%	98%
BCNN (Inception-ResnetV2, Inception-ResnetV2)	96.44%	98.66%	99%	98.33%
The from scratch “OCTorch-Net”	97.65%	96%	95.6%	95.8%