Table 2.

MRI and CT images registration performance (when the existing DNNs are used as feature extraction and coupled with a regression ANN).

Methodology	DSC	JSC	$R_p$	$R_s$	CMS	SSIM	$E_{t(s)}$
Chee and Wu20	0.9835	0.9780	0.9820	0.9815	0.9870	0.9685	9.50
Zheng et al.37	0.9865	0.9650	0.9630	0.9825	0.9730	0.9650	7.50
Miao et al.26	0.9790	0.9700	0.9610	0.9710	0.9730	0.9550	35.00
Sloan et al.33 (CNN)	0.9785	0.9780	0.9765	0.9850	0.9890	0.9530	07.35
Sloan et al.33 (FCN)	0.9895	0.9650	0.9835	0.9870	0.9855	0.9660	08.85
Liu et al.29	0.9610	0.9385	0.9670	0.9790	0.9700	0.9385	38.50
Zou et al.27	0.9680	0.9565	0.9835	0.9880	0.9785	0.9500	12.80
Proposed	0.9910	0.9820	0.9903	0.9890	0.9930	0.9700	02.50

Best result for each metric is shown in bold.