No-Reference Stereo Image Quality Assessment of Cyclopean Images Optimized Using Quaternion Wavelet Transform

Yifan Li; Chaofeng Li; Qingbing Sang

doi:10.3788/LOP56.181006

Journals >Laser & Optoelectronics Progress >Volume 56 >Issue 18 >Page 181006 > Article

Laser & Optoelectronics Progress
Vol. 56, Issue 18, 181006 (2019)

No-Reference Stereo Image Quality Assessment of Cyclopean Images Optimized Using Quaternion Wavelet Transform

Yifan Li¹, Chaofeng Li^2、*, and Qingbing Sang^1、**

Author Affiliations

¹ College of Internet of Things Engineering, Jiangnan University, Wuxi, Jiangsu 214166, China

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

show less

DOI: 10.3788/LOP56.181006 Cite this Article Set citation alerts

Yifan Li, Chaofeng Li, Qingbing Sang. No-Reference Stereo Image Quality Assessment of Cyclopean Images Optimized Using Quaternion Wavelet Transform[J]. Laser & Optoelectronics Progress, 2019, 56(18): 181006 Copy Citation Text

show less

Fig. 1. Decomposition of image Lena by first-level quaternion wavelet transform

Download full size

Fig. 2. Cyclopean images produced by proposed method. (a) Left view; (b) right view; (c) cyclopean image

Download full size

Fig. 3. Histogram of MSCN coefficient of cyclopean image

Download full size

Fig. 4. Histogram of horizontal MSCN coefficient of cyclopean image

Download full size

Fig. 5. Flowchart of proposed algorithm

Download full size

Method	WN	JP2K	JPEG	Blur	FF	All
Method in Ref. [3]	0.942	0.912	0.603	0.942	0.776	0.917
Method in Ref. [4]	0.917	0.907	0.695	0.917	0.735	0.895
Method in Ref. [6]	0.961	0.917	0.746	0.953	0.721	0.922
Method in Ref. [8]	0.923	0.898	0.632	0.928	0.845	0.926
Method in Ref. [9]	0.925	0.848	0.626	0.899	0.707	0.887
Method in Ref. [7]	0.919	0.938	0.806	0.881	0.758	0.917
Proposed method	0.963	0.935	0.682	0.958	0.877	0.947

Table 1. Comparison of PLCC on LIVE3D Phase I

Method	WN	JP2K	JPEG	Blur	FF	All
Method in Ref. [3]	0.948	0.888	0.530	0.925	0.707	0.916
Method in Ref. [4]	0.919	0.863	0.617	0.878	0.652	0.891
Method in Ref. [6]	0.906	0.838	0.603	0.791	0.679	0.903
Method in Ref. [8]	0.909	0.889	0.613	0.877	0.758	0.925
Method in Ref. [9]	0.931	0.837	0.638	0.833	0.649	0.892
Method in Ref. [7]	0.910	0.917	0.782	0.865	0.666	0.911
Proposed method	0.945	0.892	0.667	0.927	0.829	0.928

Table 2. Comparison of SROCC on LIVE3D Phase I

Method	WN	JP2K	JPEG	Blur	FF	All
Method in Ref. [3]	5.581	5.320	5.216	4.822	7.837	6.533
Method in Ref. [4]	6.433	5.402	4.523	5.898	8.322	7.247
Method in Ref. [6]	3.931	4.873	3.959	4.326	8.613	6.258
Method in Ref. [8]	6.107	6.000	5.926	6.839	6.135	6.148
Method in Ref. [7]	6.664	4.943	4.391	6.938	9.317	6.598
Proposed method	3.506	3.645	3.308	3.081	4.993	5.032

Table 3. Comparison of RMSE on LIVE3D Phase I

Method	WN	JP2K	JPEG	Blur	FF	All
Method in Ref. [3]	0.957	0.834	0.862	0.963	0.901	0.900
Method in Ref. [4]	0.947	0.899	0.901	0.941	0.932	0.895
Method in Ref. [6]	0.953	0.847	0.888	0.968	0.944	0.913
Method in Ref. [9]	0.904	0.634	0.647	0.967	0.851	0.818
Method in Ref. [7]	0.920	0.867	0.829	0.878	0.836	0.845
Proposed method	0.976	0.922	0.853	0.969	0.946	0.919

Table 4. Comparison of PLCC on LIVE3D Phase II

Method	WN	JP2K	JPEG	Blur	FF	All
Method in Ref. [3]	0.940	0.814	0.843	0.908	0.884	0.889
Method in Ref. [4]	0.950	0.867	0.867	0.900	0.933	0.880
Method in Ref. [6]	0.946	0.845	0.818	0.903	0.899	0.905
Method in Ref. [9]	0.893	0.553	0.593	0.869	0.828	0.825
Method in Ref. [7]	0.932	0.864	0.839	0.846	0.860	0.888
Proposed method	0.960	0.903	0.832	0.927	0.945	0.917

Table 5. Comparison of SROCC on LIVE3D Phase II

Method	WN	JP2K	JPEG	Blur	FF	All
Method in Ref. [3]	3.368	5.562	3.865	3.747	4.966	4.987
Method in Ref. [4]	3.513	4.298	3.342	4.725	4.180	5.102
Method in Ref. [6]	3.547	5.482	4.169	4.453	4.199	4.657
Method in Ref. [7]	4.325	5.087	4.756	6.662	6.519	7.279
Proposed method	2.011	3.192	2.547	2.376	3.521	4.085