Table 3. LSSIM results for quantitative comparison of RNLM and RNLM-CPP algorithms with parameters (r_p  = 1, r_s  = 5,  = 4, and  = 5) for T1w, T2w and PDw images.

Data type	Filtering method	Noise level
		1%	3%	5%	7%	9%
T1w	No filtering	0.9979	0.9806	0.9423	0.9025	0.8747
	RNLM	0.9895	0.9865	0.9705	0.9654	0.9448
	RNLM-CPP	0.9989	0.9940	0.9810	0.9785	0.9634
T2w	No filtering	0.9993	0.9918	0.9822	0.9686	0.9472
	RNLM	0.9878	0.9853	0.9778	0.9689	0.9661
	RNLM-CPP	0.9995	0.9961	0.9932	0.9880	0.9825
PDw	No filtering	0.9992	0.9934	0.9830	0.9575	0.9437
	RNLM	0.9769	0.9737	0.9686	0.9400	0.9365
	RNLM-CPP	0.9995	0.9969	0.9956	0.9783	0.9662