Diagnosis of Retinal Diseases Based on Bayesian Optimization Deep Learning Network Using Optical Coherence Tomography Images

Table 8

Performance of Xception.

Class labels	Precision (%)		Recall (%)		F1-score (%)
	Feature extractor	Fine tuner	Feature extractor	Fine tuner	Feature extractor	Fine tuner
AMD	92.17	98.23	89.23	98.56	90.12	96.99
CNV	90.56	98.18	91.10	97.34	90.06	95.61
DME	89.91	96.99	88.97	98.13	89.97	97.03
CSR	88.65	97.26	90.45	96.99	91.22	98.41
DR	92.11	98.61	89.34	95.09	90.23	96.21
Drusen	91.99	97.49	92.31	95.01	91.24	95.24
MH	92.05	98.12	91.81	96.38	90.03	95.02
Normal	90.51	94.78	89.45	95.73	91.45	96.12
Macro average	90.99	97.46	90.33	96.65	90.54	96.33
Weighted average