We are sharing our dataset and developed tools that we hope will enable in creating more immersive virtual reality experiences. This repository is for open source codes and materials for the paper: Do users behave similarly in VR? Investigation of the influence on the system design
Each available codes in this repository is based on Python 3.
Firstly, you should install the following dependencies:
- pip install wget
A python script, named download.py
is provided to download the dataset.
The following script can download all the dataset and locate it into the project folder.
- python download.py
Alternatively, you can download it with this link.
Figure 1 illustrates the organization of the dataset. The folder of the downloaded and extracted data
folder contains three sub-folders, namely trajectories
and videos
. The trajectories
folder contains three subfolders in which each one represents a device: dev_0
: HMD ,dev_1
: Laptop, and dev_2
: Tablet. Each sub-folder is then contained a number of sub-sub-folders, omnidirectional video (ODV) folder, corresponding the names of the ODVs, e.g., v01_BabyPandas. Each ODV folder contains the number of CSV files which corresponds to the number of participants. The viewport trajectories are stored as comma-separated values (CSV) files in ASCII. The name of each CSV file specifies the participant ID, and each line of the CSV file has the following structure:
| u | v | university | category | | ------------ | ------------- | |0.523| 0.564| tcd| cat1|
u: u coordinate of the viewport center (normalized value).
v: v coordinate of the viewport (normalized value).
university: subjective experiment location.
category: category type of ODV.
The videos
folder contains three sub-folders in which each one represents a category ID number: cat1
: Documentary, cat2
: Movie , and cat3
: Action. Each sub-folder is then contained a number of sub-sub-folders, omnidirectional video (ODV) folder, corresponding the names of the ODVs, e.g., v01_BabyPandas.
_Server_Opt_Rossi_Script.py
This script is able to evaluate the proposed server optimisation for VR tile-based ODV.
_Server_Opt_Script_Netflix_Apple.py
This script is able to evaluate the Netflix and Apple optimisation for VR tile-based ODV.
prob_dev_0.csv
It contains heatmap probability per tile for the HMD device. It was estimated using viewport heatmaps. The .csv contains the following structure: | folder location | file name | probability for tile 0 | probability for tile 1 | probability for tile 2 | probability for tile 3 | probability for tile 4 |
prob_dev_1.csv
Same asprob_dev_o.csv
, but, this file is for Laptop.
prob_dev_2.csv
Same asprob_dev_o.csv
, but, this file is for HMD.
RD_1920x1080.csv
: It contains RD results for each tile for 1920x1080 resolution. The measurement is based on WS-PSNR and WS-MSE. q_i is quality for the i-th tile, d_i is distortion for the i-the tile, b_i is file size for the i-th tile.
RD_1920x1080_Apple.csv
It contains RD results for Apple optimization scenario for 1920x1080 resolution.
RD_1920x1080_Netflix.csv
It contains RD results for Netflix optimization scenario for 1920x1080 resolution.
RD_2560x1440.csv
Similar as above, but it is for 2560x1440 display resolution.
RD_2560x1440_Apple.csv
Similar as above, but it is for 2560x1440 display resolution.
RD_2560x1440_Netflix.csv
Similar as above, but it is for 2560x1440 display resolution.
RD_3840x2160.csv
Similar as above, but it is for 3840x2160 display resolution.
RD_3840x2160_Apple.csv
Similar as above, but it is for 3840x2160 display resolution.
RD_3840x2160_Netflix.csv
Similar as above, but it is for 3840x2160 display resolution.
Please cite our paper in your publications if it helps your research:
@article{rossi2020,
title = {Do users behave similarly in VR? Investigation of the influence on the system design},
author = {Silvia Rossi and Cagri Ozcinar and Aljosa Smolic and Laura Toni},
year = {2020},
booktitle = {Transactions on Multimedia Computing Communications and Applications}
}
| Silvia Rossi | Cagri Ozcinar | Aljosa Smolic | Laura Toni |
This work has been partially funded by Adobe under Academic Donation scheme. Also, this publication has emanated from research supported in part by a research grant from Science Foundation Ireland (SFI) under the Grant Number 15/RP/2776.