Table 2.

Statistical analysis of MS-RDN and RED-CNN reconstructions using generalized linear models.

	NMSE (dB)	Bias (${\text{cm}}^{-1}$)	PSNR (dB)	SSIM
Single slice training, non-TOI	$+0.034$	$-0.399\times {10}^{-4}$	$+0.034$	$+0.0003$
Single slice training	$+0.197$	$-1.501\times {10}^{-4}$	$+0.197$	$+0.0006$
Multi-slice training ($Z=5$)	$+0.144$	$-0.975\times {10}^{-4}$	$+0.144$	$+0.0007$

The table reports the performance gained by MS-RDN over RED-CNN for three different configurations and four quantitative metrics.

All improvements are significant with $P<0.0001$. Please note that the values corresponding to NMSE and PSNR are identical since these quantities are related as shown in Eqs. (6) and (7).