EURASIP Journal on Image and Video Processing

Table 1 Performances of proposed metric and competitors on three benchmark databases in terms of SROCC, PLCC and RMSE

From: Phase congruency based on derivatives of circular symmetric Gaussian function: an efficient feature map for image quality assessment

		SROCC				PLCC				RMSE
		LIVE	CSIQ	TID-2013	Ave	LIVE	CSIQ	TID-2013	Ave	LIVE	CSIQ	TID-2013
Grayscale	PSNR	0.8756	0.8058	0.6394	0.7100	0.8723	0.7512	0.7017	0.7395	13.3597	0.1733	12.2420
	SSIM	0.9479	0.8756	0.7417	0.8012	0.9449	0.8613	0.7895	0.8289	8.9455	0.1334	10.5462
	MS-SSIM	0.9513	0.9133	0.7859	0.8374	0.9489	0.8991	0.8329	0.8647	8.6188	0.1149	9.5098
	IW-SSIM	0.9567	0.9213	0.7779	0.8346	0.9522	0.9144	0.8319	0.8675	8.3472	0.1063	9.5364
	RFSIM	0.9434	0.9291	0.7743	0.8317	0.9386	0.9164	0.8329	0.8662	9.4298	0.1051	9.5089
	IFC	0.9259	0.7671	0.5390	0.6463	0.9268	0.8366	0.7220	0.7777	10.2641	0.1438	11.8900
	VIF	0.9636	0.9195	0.6770	0.7703	0.9604	0.9277	0.7720	0.8326	7.6137	0.0980	10.9215
	FSIM	0.9634	0.9240	0.8015	0.8515	0.9597	0.9120	0.8589	0.8857	7.6780	0.1077	8.8003
	GMSD	0.9603	0.9570	0.8044	0.8590	0.9603	0.9541	0.8590	0.8937	7.6214	0.0786	8.7966
	\({q}_{m,g}\)	0.9508	0.9193	0.7856	0.8382	0.9466	0.9010	0.8370	0.8673	8.8069	0.1139	9.4030
	\({q}_{sd,g}\)	0.9579	0.9494	0.8101	0.8609	0.9534	0.9453	0.8746	0.9010	8.2420	0.0857	8.3324
Color	FSIMc	0.9645	0.9310	0.8510	0.8849	0.9613	0.9192	0.8769	0.8989	7.5296	0.1034	8.2600
	\({q}_{m,c}\)	0.9508	0.9270	0.8370	0.8729	0.9466	0.9088	0.8507	0.8776	8.8100	0.1095	9.0338
	\({q}_{sd,c}\)	0.9583	0.9506	0.8407	0.8809	0.9533	0.9467	0.8672	0.8965	8.2487	0.0845	8.5558
Deep-learning	DISTS-Gray	0.942	0.905	0.764	0.820	–	–	–	–	–	–	–
	DISTS-Color	0.954	0.929	0.830	0.869	–	–	–	–	–	–	–
	DeepSim	0.974	0.919	0.846	0.881	0.968	0.919	0.872	0.897	-	-	-

The preferable values of conventional metrics are shown in boldface for each database

Back to article page