Table 4.

The resulting precision, recall, and f1-score values for Scenario 1 (training the algorithms with all training datasets and testing them on the Taiwan, China, Japan, and all test datasets separately).

		U-Net			ResU-Net
Size	32	64	128	32	64	128
Precision	67.45	63.95	64.32	66.39	69.65	60.11
Recall	58.97	67.13	61.39	6543	63.85	74.29
F1-score	62.93	65.5	62.82	65.91	66.63	66.45
Precision	65.47	60.48	59.73	64.99	68.11	57.21
Recall	77.52	84.16	80.24	78.94	78.51	88.38
F1-score	70.99	70.38	68.48	71.29	72.94	69.46
Precision	79.76	70.27	81.2	75.02	74.74	65.57
Recall	55.52	61.85	55.48	59.04	60.33	64.3
F1-score	65.47	65.79	65.92	66.07	66.76	64.93
Precision	69.26	68.35	70.8	67.33	71.12	63.41
Recall	44.05	53.62	46.33	54.92	52.14	63.43
F1-score	53.85	60.1	56.01	60.49	60.17	63.42
Algorithms trained by sample from			All	Taiwan	China	Japan

The highest values of precision, recall, and f1-score are indicated in bold.