Quality Assessment Without Reference Images Based on Convolution Neural Network and Deep Forest

Yindong Chen; Chaofeng Li; Qingbing Sang

doi:10.3788/LOP56.111003

Journals >Laser & Optoelectronics Progress >Volume 56 >Issue 11 >Page 111003 > Article

Laser & Optoelectronics Progress
Vol. 56, Issue 11, 111003 (2019)

Quality Assessment Without Reference Images Based on Convolution Neural Network and Deep Forest

Yindong Chen¹, Chaofeng Li², and Qingbing Sang^1、*

Author Affiliations

¹ School of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214122, China

² Institute of Logistics Science & Engineering, Shanghai Maritime University, Shanghai 200135, China;

show less

DOI: 10.3788/LOP56.111003 Cite this Article Set citation alerts

Yindong Chen, Chaofeng Li, Qingbing Sang. Quality Assessment Without Reference Images Based on Convolution Neural Network and Deep Forest[J]. Laser & Optoelectronics Progress, 2019, 56(11): 111003 Copy Citation Text

show less

Fig. 1. Architecture of CNN

Download full size

Fig. 2. Architecture of deep regression forest

Download full size

Fig. 3. Overall framework of proposed model

Download full size

Fig. 3. [in Chinese]

Download full size

Fig. 4. Predicted scatter plot of TID2008 database

Download full size

Imagedatabase	Number ofreferenceimages	Number ofdistortedimages	Number ofdistortiontypes	Numberof peopletested
LIVE	29	779	5	161
TID2008	25	1700	17	838

Table 1. Standard databases of image quality assessment

Model	JP2K	JPEG	WN	BLUR	FF	ALL
DIIVINE^[6]	0.913	0.910	0.984	0.921	0.863	0.916
BLIINDS-II^[20]	0.929	0.942	0.969	0.923	0.889	0.931
BRISQUE^[7]	0.914	0.965	0.979	0.951	0.877	0.940
CORNIA^[21]	0.943	0.955	0.976	0.969	0.906	0.942
CNN^[11]	0.952	0.977	0.978	0.962	0.908	0.956
Proposed	0.967	0.974	0.983	0.971	0.938	0.972

Table 2. Comparison of SROCC experimental results on LIVE database

Model	JP2K	JPEG	WN	BLUR	FF	ALL
DIIVINE^[6]	0.922	0.921	0.988	0.923	0.888	0.917
BLIINDS-II^[20]	0.935	0.968	0.980	0.938	0.896	0.930
BRISQUE^[7]	0.922	0.973	0.985	0.951	0.903	0.942
CORNIA^[21]	0.951	0.965	0.987	0.968	0.917	0.935
CNN^[11]	0.953	0.981	0.984	0.953	0.933	0.953
Proposed	0.971	0.979	0.983	0.985	0.944	0.975

Table 3. Comparison of LCC experimental results on LIVE database

Model	SROCC	LCC
CORNIA^[21]	0.890	0.880
BRISQUE^[7]	0.882	0.892
CNN^[11]	0.920	0.903
Proposed	0.935	0.929

Table 4. Comparison of SROCC and PLCC experimental results on TID2008 database

Yindong Chen, Chaofeng Li, Qingbing Sang. Quality Assessment Without Reference Images Based on Convolution Neural Network and Deep Forest[J]. Laser & Optoelectronics Progress, 2019, 56(11): 111003

Download Citation

Set citation alerts for the article

Tools

Set citation alerts for the article

Save the article for my favorites

Paper Information