Table 6.

The resulting precision, recall, and f1-score values for Scenario 3 (training the algorithms with the training dataset of China and testing on the Taiwan, China, Japan, and all test datasets separately).

		U-Net			ResU-Net
Size	32	64	128	32	64	128
Precision	58.61	46.68	51.98	47.19	43.33	42.42
Recall	56.57	60.75	59.69	61.75	53.35	59.37
F1-score	57.57	52.8	55.57	53.5	47.85	49.48
Precision	50.31	37.95	43.17	35.95	29.36	34.09
Recall	81.25	90.9	87.62	75.31	57.31	89.92
F1-score	62.14	53.55	57.84	48.67	38.83	49.44
Precision	78.59	75.52	81.25	72.25	75.91	71.74
Recall	63.55	64.66	59.58	71.09	69.92	65.4
F1-score	70.28	69.67	68.75	71.7	72.9	68.48
Precision	79.98	80.32	81.03	71.6	7312	81.17
Recall	35.66	35.64	36.76	50.05	48.46	33.84
F1-score	49.33	49.37	50.58	58.92	58.29	47.76
Algorithms trained by sample from			All	Taiwan	China	Japan

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