Remote Sensing Image Change Detection Network Based on Twin High-Resolution Representation

Table 3

Results of the quantitative evaluation of the different methods on the CDD and DSIFN datasets.

Models	CDD				DSIFN
	Precision (%)	Recall (%)	F1 (%)	OA (%)	Precision (%)	Recall (%)	F1 (%)	OA (%)
FCN-PP	0.8264	0.8060	0.8047	0.9536	0.5640	0.6703	0.6126	0.8559
FC-siam-conc	0.8441	0.8250	0.8250	0.9572	0.4183	0.5963	0.4917	0.7905
FC-siam-diff	0.8578	0.8364	0.8373	0.9575	0.5151	0.6554	0.5769	0.8366
Unet++_MSOF	0.8954	0.8711	0.8756	0.9673	0.5983	0.6591	0.6273	0.8668
IFN	0.9496	0.8608	0.9030	0.9771	0.6711	0.6754	0.6733	0.8886
TCANet (ours)	0.9670	0.8798	0.9213	0.9788	0.9104	0.8893	0.8987	0.9537

The best ones are marked in bold.