Descriptions:

The new Waterloo-IVC 3D Video Quality Database Phase I is created from 4 pristine multi-view 3D videos, i.e., Balloons, Book, Kendo, and Lovebird, which are commonly used 3D HEVC testing sequences. The new Waterloo-IVC 3D Video Quality Database Phase II is created from 6 pristine stereoscopic 3D videos, i.e., Barrier, Craft, Laboratory, Soccer, Tree, and Dancer, which were collected from previous subjective 3D video quality studies [1], [2] and [3].

The Waterloo-IVC 3D Video database Phase I includes stereoscopic 3D videos obtained from symmetric and asymmetric transform-domain quantization coding followed by different levels of low-pass filtering. Each single-view video was compressed using an HEVC encoder by five levels of transform-domain quantization with QP = {25, 35, 40, 45, 50} in low-delay main profile. The single-view videos were employed to generate compressed stereoscopic videos, either symmetrically or asymmetrically.

The Waterloo-IVC 3D Video database Phase II includes various stereoscopic 3D videos obtained from mixed-resolution coding, asymmetric transform-domain quantization coding, their combinations, and different levels of low-pass filtering. Three choices of pre-processing, i.e., pre-downsampling by 2, pre downsampling by 4, and pre-processed Gaussian lowpass filtering with σ = 2.5, are applied to each single-view video. Then the single-view video was compressed using an HEVC encoder by different levels of transform-domain quantization with QP = {25, 30, 35, 40, 45} in low-delay main profile. The single-view videos were employed to generate compressed stereoscopic videos after upsampling if needed, either symmetrically or asymmetrically.

Publications:

• Jiheng Wang, Shiqi Wang and Zhou Wang, "Asymmetrically compressed stereoscopic 3D videos: quality assessment and rate-distortion performance evaluation", IEEE Transactions on Image Processing, vol. 26, no. 3, pp. 1330-1343, Mar. 2017. [pdf]
• Jiheng Wang, Shiqi Wang and Zhou Wang, "Quality prediction of asymmetrically compressed stereoscopic videos", IEEE International Conference on Image Processing, Québec City, Québec, Canada, Sept. 27-30, 2015. [pdf][poster]

References:

[1] M. Urvoy, M. Barkowsky, R. Cousseau, Y. Koudota, V. Ricorde, P. Le Callet, J. Gutiérrez, and N. García, “NAMA3DS1-COSPAD1: Subjective video quality assessment database on coding conditions introducing freely available high quality 3D stereoscopic sequences”, International Workshop on Quality of Multimedia Experience, Yarra Valley, VIC, Australia, July 5-7, 2012.

[2] M. H. Pinson, “The consumer digital video library [best of the web]”, IEEE Signal Processing Magazine, vol. 30, no. 4, pp. 172–174, July 2013.

[3] ISO/IEC JTC1/SC29/WG11, “Undo Dancer 3DV sequence for purposes of 3DV standardization,” Document M20028, ISO, Geneva, Switzerland, 2011.

Download Now: MOS [5.07MB], Matlab Code [6.90KB], Phase I [6.37GB] and Phase II [40.20GB]

Password: uwivc3D2016

Source code for the proposed 2D-to-3D quality prediction model is also available! Please contact me at j237wang@gmail.com.