8000 GitHub - gfxdisp/AR-DAVID: Repository for AR-DAVID (Siggraph Asia 2024)
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

gfxdisp/AR-DAVID

Repository files navigation

AR-DAVID: Augmented Reality Display Artifact Video Dataset

Alexandre Chapiro Dongyeon Kim, Yuta Asano, and Rafał K. Mantiuk

Source code for the SIGGRAPH Asia 2024 paper titled "AR-DAVID: Augmented Reality Display Artifact Video Dataset"

Get Started

Create anaconda environment. Our code has been implemented and tested on Windows.

conda create -n ardavid python=3.10
conda activate ardavid

1. Add the Evaluation Module

Add the cvvdp module as a Git submodule:

git submodule add https://github.com/gfxdisp/ColorVideoVDP.git

2. Install the ColorVideoVDP

Follow the installation instructions provided in the ColorVideoVDP repository.

Make sure to install any dependencies and build requirements mentioned there.


3. Download the AR-DAVID Dataset

Download the dataset from the following link:

🔗 AR-DAVID Dataset

3.1. Place or Link the Dataset

You have two options:

  • Move the dataset to:

    datasets/AR-DAVID
    
  • OR, optionally create a symbolic link to the dataset directory:

    On Linux/macOS:

    ln -s /path/to/AR-DAVID datasets/AR-DAVID

4. Install Additional Libraries

Install the required Python additional libraries.


5. Test

  • Different optical blending methods

    python run_metric.py --metric=cvvdp --fusion-method={fusion_method}

    Fusion_method can be none,mean,pinhole,pinhole-stereo,blur,blur-stereo.

  • Background discounting

    python run_metric.py --metric=cvvdp --fusion-method={fusion_method} --discount_factor={d}

    The value for d can be in a range from 0 to 1.

  • For debugging

    python run_metric.py --metric=dm-preview --fusion-method={fusion_method} --discount_factor={d}

    This mode will save the processed videos in dataset/AR-DAVID-{additional_method}/preview/.


6. Plot

This is a simple MATLAB example demonstrating how to generate some of the figures in the main paper.

📁 Files

  • analysis/plot_across_backgrounds.m: A script that generates the scaled quality per scene. Colors represent the distortion type, while line styles indicate the strength of the distortion levels.

  • analysis/plot_scatter_backgrounds.m: A script that generates a scatter plot for each optical blending method. Colors represent background luminance levels, and markers indicate the background type.

🚀 How to Run

  1. Open MATLAB.

  2. Navigate to the folder using the command window or cd:

    cd path/to/this/folder/analysis
  3. Run the script:

    plot_across_backgrounds
    plot_scatter_backgrounds

Citation

@article{chapiro2024ar,
title={AR-DAVID: Augmented Reality Display Artifact Video Dataset},
author={Chapiro, Alexandre and Kim, Dongyeon and Asano, Yuta and Mantiuk, Rafa{\l} K},
journal={ACM Transactions on Graphics (TOG)},
volume={43},
number={6},
articleno={186},
pages={1--11},
year={2024},
publisher={ACM New York, NY, USA}
}

Contact

If you have any questions, please contact

Releases

No releases published

Packages

No packages published
0