Quality Assessment for Volumetric Video Compression

12th November 2018
Quality Assessment for Volumetric Video Compression

Volumetric video (VV) pipelines reached a high level of maturity, creating interest to use such content in interactive visualization scenarios. Volumetric video allows real-world content to be captured and represented as 3D models, which can be viewed from any chosen viewpoint and direction. Thus, it is ideal to be used in augmented reality (AR) or virtual reality (VR) applications. Both textured polygonal meshes and point clouds are popular methods to represent volumetric videos. We study both subjective and objective quality of the volumetric videos for one of the most common application scenarios: compression.

On this page, we share two different volumetric video quality databases hoping that these databases will be useful for other researchers working on quality assessment for volumetric videos. These databases are named as vsenseVVDB and vsenseVVDB2 and are explained below.

V-SENSE Volumetric Video Quality Database #1

The vsenseVVDB database comprised of two different contents (Matis and Rafa, see below) where human actors play with a football. These volumetric videos are represented as point clouds and are compressed using a state-of-the-art compression algorithm: MPEG Point Cloud Compression Test Model Category 2 (TMC2) version 1, which is renamed as V-PCC. To the best of our knowledge, this study is the first to consider TMC2 compression for volumetric video represented as colored point clouds and study its effects on the perceived quality. Here we provide the reference colored point clouds (PLY) and compressed bitstreams, along with subjective quality scores.

V-SENSE Volumetric Video Quality Database #2

The vsenseVVDB2 database comprised of eight different contents with different human actors. Four of these eight contents are from V-SENSE (AxeGuy, LubnaFriends, Rafa2, Matis, from left to right below) and the remaining four are from the 8i database (Soldier, Redandblack, Loot, Longdress, from right to left below). As we do not have the right to redistribute 8i point clouds, we ask interested readers to download the database from their own website. Here, we provide the four generated by V-SENSE in both textured mesh (OBJ) and colored point cloud (PLY) formats, along with the compressed bitstreams, reconstructed models, and the subjective quality scores.

DOWNLOADS

The vsenseVVDB (Volumetric Video Quality Database #1) can be downloaded here: [vsenseVVDB dataset].

The vsenseVVDB2 (Volumetric Video Quality Database #2) can be downloaded here: [vsenseVVDB2 Database] [vsenseVVDB2 lite version (subjective scores only)]

 

Please kindly cite our papers in your publication if you use one of these databases:

  • For vsenseVVDB:
  • E. Zerman, P. Gao, C. Ozcinar, A. Smolic. “Subjective and objective quality assessment for volumetric video compression.” IS&T Electronic Imaging, Image Quality and System Performance XVI, San Francisco, California, USA, January 2019.
@inproceedings{zerman2019subjective, 
  title     = {Subjective and Objective Quality Assessment for Volumetric Video Compression}, 
  author    = {Zerman, Emin and Gao, Pan and Ozcinar, Cagri and Smolic, Aljosa}, 
  year      = {2019}, 
  booktitle = {{IS\&T} Electronic Imaging, Image Quality and System Performance {XVI}} 
}
  • For vsenseVVDB2:
  • E. Zerman, C. Ozcinar, P. Gao, A. Smolic. “Textured mesh vs coloured point cloud: A subjective study for volumetric video compression.” Twelfth International Conference on Quality of Multimedia Experience (QoMEX), Athlone, Ireland, May 2020.
@inproceedings{zerman2020textured,
  title     = {Textured mesh vs coloured point cloud: A subjective study for volumetric video compression},
  author    = {Zerman, Emin and Ozcinar, Cagri and Gao, Pan and Smolic, Aljosa},
  year      = {2020},
  booktitle = {Twelfth International Conference on Quality of Multimedia Experience ({QoMEX})}
}

ACKNOWLEDGEMENT

This publication has emanated from research conducted with the financial support of Science Foundation Ireland (SFI) under the Grant Number 15/RP/2776.

CONTACT

If you have any question, send an e-mail to zermane@scss.tcd.ie or ozcinarc@scss.tcd.ie or gaopa@scss.tcd.ie.

2020

Zerman, Emin; Ozcinar, Cagri; Gao, Pan; Smolic, Aljosa

Textured Mesh vs Coloured Point Cloud: A Subjective Study for Volumetric Video Compression Inproceedings Forthcoming

Twelfth International Conference on Quality of Multimedia Experience (QoMEX), IEEE Athlone, Ireland, Forthcoming.

Abstract | Links | BibTeX

2019

Zerman, Emin; Valenzise, Giuseppe; Smolic, Aljosa

Analysing the Impact of Cross-Content Pairs on Pairwise Comparison Scaling Inproceedings

11th International Conference on Quality of Multimedia Experience (QoMEX 2019), IEEE 2019.

Abstract | Links | BibTeX

Zerman, Emin; Gao, Pan; Ozcinar, Cagri; Smolic, Aljosa

Subjective and Objective Quality Assessment for Volumetric Video Compression Inproceedings

IS&T Electronic Imaging, Image Quality and System Performance XVI, 2019.

Abstract | Links | BibTeX