TABLE 1.

Performance comparison of the OIQA database.

Models	JPEG			JP2K			WN			BLUR			ALL
	PLCC	SRCC	RMSE	PLCC	SRCC	RMSE	PLCC	SRCC	RMSE	PLCC	SRCC	RMSE	PLCC	SRCC	RMSE
BRISQUE (Mittal et al., 2012)	0.9401	0.8691	0.7605	0.8986	0.8447	0.9009	0.9663	0.9162	0.4664	0.9163	0.8529	0.7295	0.8972	0.8694	0.9358
DIIVINE (Moorthy and Bovik, 2011)	0.8963	0.8132	0.9307	0.9337	0.8970	0.7829	0.9649	0.9011	0.5029	0.9205	0.8392	0.7413	0.8793	0.8458	1.0127
SSEQ (Liu et al., 2014)	0.8905	0.8194	0.9945	0.8900	0.8500	0.9656	0.9589	0.9103	0.5200	0.9508	0.8989	0.6030	0.8970	0.8750	0.9240
OG-IQA (Liu et al., 2016)	0.9552	0.8912	0.6845	0.8759	0.8253	1.0224	0.9717	0.9206	0.4262	0.9473	0.9025	0.6005	0.9076	0.8954	0.8684
NRSL (Li et al., 2016)	0.9490	0.8834	0.7260	0.9538	0.8941	0.5507	0.9176	0.8691	0.8370	0.9258	0.8618	0.7074	0.8852	0.8537	0.9749
SSP-BOIQA (Zheng et al., 2020)	0.877	0.834	–	0.853	0.852	–	0.905	0.843	–	0.854	0.862	–	0.860	0.865	–
MC360IQA (Sun et al., 2019)	0.9015	0.8995	0.8234	0.8861	0.8779	1.3687	0.9195	0.9124	0.8234	0.8938	0.8892	1.3838	0.8953	0.8928	1.5052
Zhou Y. et al. (2022)	0.936	0.940	–	0.920	0.934	–	0.968	0.957	–	0.925	0.920	–	0.899	0.923	–
S3DAVS	0.9267	0.8872	0.8634	0.9370	0.9306	0.7717	0.9725	0.9623	0.4384	0.9692	0.9662	0.4805	0.9405	0.9348	0.7183

The best-performing NR metrics are highlighted in bold.