VQA-ODV

A large-scale VQA dataset of omnidirectional video proposed in ACMMM 2018 paper.

Chen Li, Mai Xu*, Xinzhe Du, Zulin Wang. 2018. Bridge the Gap Between VQA and Human Behavior on Omnidirectional Video: A Large-Scale Dataset and a Deep Learning Model. In 2018 ACM Multimedia Conference (MM ’18), October 22–26, 2018, Seoul, Republic of Korea. ACM, New York, NY, USA, 9 pages. https://doi.org/10.1145/3240508.3240581

Citation

If you use VQA-ODV in your research, please cite:

@inproceedings{Li:2018:BGV:3240508.3240581,
 author = {Li, Chen and Xu, Mai and Du, Xinzhe and Wang, Zulin},
 title = {Bridge the Gap Between VQA and Human Behavior on Omnidirectional Video: A Large-Scale Dataset and a Deep Learning Model},
 booktitle = {Proceedings of the 26th ACM International Conference on Multimedia},
 series = {MM '18},
 year = {2018},
 isbn = {978-1-4503-5665-7},
 location = {Seoul, Republic of Korea},
 pages = {932--940},
 numpages = {9},
 url = {http://doi.acm.org/10.1145/3240508.3240581},
 doi = {10.1145/3240508.3240581},
 acmid = {3240581},
 publisher = {ACM},
 address = {New York, NY, USA},
 keywords = {human behavior, omnidirectional video, visual quality assessment},

Introduction

VQA-ODV is a large-scale dataset of omnidirectional video for visual quality assessment (VQA). There is in total 600 omnidirectional video sequences in VQA-ODV, of which 60 are reference sequences in diverse content. The dataset is organized as follow. More details can be found in our ACMMM 2018 paper.

Copyright

We do not own the copyright of the video sequences. G7PressConference,G7Shooting,G8Pagoda,G8Salon,G9ConcertLive,G9DrivingInCity,G9FootballMatch,G9GreatWall,G10BoatInPark, G10BuddhaCave,G10DrivingInCountry,G10XiaoGuang are provided and owned by Letin VR Digital Technology Co., Ltd. Others are collected from YouTube and owned by their original uploaders. For researchers and educators who wish to use the video sequences for non-commercial research and/or educational purposes, we can provide access under certain conditions and terms.

The following uses are allowed for the proposed sequences:

1. Can be published in technical papers, played at technology research and development events.
2. Can be used by Standards committees. (e.g., SMPTE, ITU, MPEG, VQEG).

The following uses are NOT allowed for the proposed sequences:

1. Do not publish snapshots in product brochures.
2. Do not use video for marketing purposes
3. Do not redistribute video with a commercial product.
4. Do not use in television shows, commercials, or movies.

Download

Please send email in the format as follow:

• SUBJECT: VQA-ODV
• TO: dataset[at]buaamc2[dot]net

Or you can use this hyperlink to send the email. After sending the email, you are supposed to receive the response immediately with the download link and the password for unzipping. If not, please check whether it is identified as spam by your ESP.