Hybrid Low-Rank Tensor CP and Tucker Decomposition with Total Variation Regularization for HSI Noise Removal

Table 2

All the algorithms compare the performance metrics for PAC data sets under different noise types.

Noise type	Index	Noisy	LRTV	WTV-HMCT	BM4D	WGLRTD	KBR	LLRT
Case 1	MPSNR	28.147	32.521	38.924	37.709	39.460	40.911	39.033
	MSSIM	0.838	0.933	0.985	0.975	0.983	0.989	0.983
	MFSIM	0.909	0.953	0.990	0.981	0.989	0.991	0.989
	SAM	0.263	0.129	0.071	0.087	0.058	0.050	0.046
Case 2	MPSNR	14.163	24.510	29.275	28.179	30.290	30.646	30.489
	MSSIM	0.219	0.659	0.864	0.834	0.887	0.906	0.893
	MFSIM	0.584	0.791	0.910	0.881	0.930	0.939	0.930
	SAM	0.773	0.201	0.153	0.140	0.147	0.081	0.079
Case 3	MPSNR	8.135	21.671	25.479	24.453	26.484	26.898	27.164
	MSSIM	0.065	0.451	0.710	0.663	0.768	0.785	0.792
	MFSIM	0.405	0.683	0.819	0.801	0.863	0.870	0.859
	SAM	1.060	0.232	0.195	0.154	0.221	0.095	0.086