Table 2.

Quantitative evaluation (MEAN ± SD) for 400 test images from the RS dataset. SD denotes the standard deviation of a single metric across the test set. As explained in Section 3.6, the tissue-wise metrics (i.e., ${\mathrm{MRE}}_{\mathrm{m}}$ and ${\mathrm{PSNR}}_{\mathrm{m}}$) refer to the result within a specific tissue region. The tSD refers to the standard deviation across the mean values of tissue-wise metrics of a given method. The tMRE and tPSNR are used as image-wise metrics. The regions corresponding to the labels “bk” and “#1”–“#5” can be found in Fig. 2.

	Algorithms	Tissue-wise metrics						tSD	tMRE (top) tPSNR (bottom)
		bk	#1	#2	#3	#4	#5
$\left|\text{Relative error}\right|$	Unet-MSE	0.058 ± 0.013	0.007 ± 0.003	0.009 ± 0.003	0.031 ± 0.009	0.054 ± 0.026	0.064 ± 0.036	0.023	0.037 ± 0.009
	Unet-MSRE	0.034 ± 0.007	0.022 ± 0.006	0.023 ± 0.009	0.026 ± 0.012	0.033 ± 0.015	0.052 ± 0.029	0.010	0.032 ± 0.007
	Ours	0.014 ± 0.006	0.016 ± 0.005	0.016 ± 0.005	0.017 ± 0.006	0.018 ± 0.007	0.021 ± 0.012	0.002	0.017 ± 0.005
PSNR (dB)	Unet-MSE	22.437 ± 1.731	40.498 ± 2.513	38.268 ± 2.419	28.673 ± 2.106	24.571 ± 3.148	24.338 ± 4.507	7.057	29.797 ± 1.428
	Unet-MSRE	27.284 ± 1.654	31.009 ± 1.995	31.230 ± 2.344	30.097 ± 2.587	28.567 ± 2.917	25.749 ± 4.064	2.010	28.989 ± 1.318
	Ours	35.517 ± 2.528	34.005 ± 2.273	33.891 ± 2.137	33.702 ± 2.255	33.241 ± 2.595	33.611 ± 3.999	0.722	33.994 ± 1.703