We established a synthesized video quality database with texture/depth compression distortion and the associated subjective. Then, we proposed a full reference objective video quality assessment (VQA) algorithm primarily focusing on the temporal flicker distortions of the synthesized video, which is based on two quality features extracted from both spatial and temporal domains of the synthesized sequence. Experimental results show that the performance of the proposed VQA is significantly superior to the others for ALL DATA set, and is particularly prominent on the subsets which have significant temporal flicker distortion induced by depth compression and view synthesis process.

Provide a 3D video database and a better 3D video quality metric

Significantly improve the 3D coding efficiency, i.e. min(D), s.t. R< RT

No publicly available database for 3D synthesized video;

The distortions of 3D synthesized video are underestimated or overestimated;

Flicker artifacts in synthesized video;

Subjective evaluation is expensive and time-consuming;

3D VQA is required for human 3D perception and video system optimization.

We established a 3D synthesized video quality database and proposed a full reference VQA based on two key features of 3D video. The performance is significantly superior to the state-of-the-art schemes.

Subjective quality assessment

Basic idea: the flickering can be seen as the high frequency direction variation.

The First 3D synthesized video database for quality assessment in worldwide; Ten different MVD sequences, 140 synthesized videos, 40 subjective scores, DMOS are provided.
http://codec.siat.ac.cn/SIAT_Synthesized_Video_Quality_Database/index.html

Objective quality assessment

Basic idea: the metric is based on the temporal flicker feature and the spatio-temporal activity feature

Our metric significantly improves the PLCC , SROCC, and RMSE as compared with state-of-the-art video/image quality metrics.

Our metric VQA_SIAT is more consistent with human perception

X. Liu, Y. Zhang, S. Hu, S. Kwong, C.C. Jay Kuo, and Q. Peng, Subjective and Objective Video Quality Assessment of 3-D Synthesized View with Texture/Depth Compression Distortion, IEEE Transactions on Image Processing ,vol.24, no.12, pp4847-4861, Dec. 2015,

Y. Zhang, X. Yang, X. Liu, Y.B. Zhang, G. Jiang, and S. Kwong, High Efficiency 3D Depth Coding Based on Perceptual Video Quality of Synthesized View, IEEE Transactions on Image Processing, Sept. 2016. DOI:10.1109/TIP.2016.2615290 (SCI IF 3.735)