Table 8.

Model performance analysis.

Model	Datasets	Precision/%	Recall/%	F1-Sorce/%
ResNet50	Pu-erh raw tea face	94.59	96.16	95.37
	Pu-erh ripe tea face	91.54	95.66	93.55
	Mixed tea face	95.46	96.41	95.93
VGG16	Pu-erh raw tea face	78.37	91.83	84.57
	Pu-erh ripe tea face	92.84	93.00	92.92
	Mixed tea face	92.57	92.49	92.49
Inception-Resnet-v1	Pu-erh raw tea face	95.65	95.50	95.57
	Pu-erh ripe tea face	93.92	87.66	90.68
	Mixed tea face	92.64	96.58	94.57
MobileNet	Pu-erh raw tea face	94.39	92.66	93.52
	Pu-erh ripe tea face	96.39	93.66	95.01
	Mixed tea face	96.16	98.33	97.23
MobileNetV3	Pu-erh raw tea face	94.05	97.50	95.74
	Pu-erh ripe tea face	97.07	94.16	95.59
	Mixed tea fac	97.82	97.50	97.66
TeaFaceNet	Pu-erh raw tea face	97.34	97.83	97.58
	Pu-erh ripe tea face	98.98	97.16	98.06
	Mixed tea face	98.82	97.58	98.20

Significant are in value [bold].