Table 2.

Performance of different models for independent test sets. Nos. 1 to 6 represent the 6 groups of established models. No. 7 to No. 11 represent 5 RIFSeg-Nets constructed separately using different backbones.

No.	Methods	P	R	F1-score
1	FCN	0.85	0.77	0.81
2	U-net	0.88	0.82	0.85
3	SegNet	0.88	0.85	0.86
4	FuseNet	0.86	0.81	0.83
5	MFNet	0.91	0.87	0.89
6	PSPNet	0.91	0.91	0.91
7	RTFNet	0.92	0.90	0.91
8	RIFSeg-Net-18	0.90	0.86	0.88
9	RIFSeg-Net-34	0.89	0.88	0.88
10	RIFSeg-Net-50	0.94	0.93	0.93
11	RIFSeg-Net-101	0.92	0.91	0.91
12	RIFSeg-Net-152	0.93	0.91	0.92