Hole-robust Wireframe Detection

Data and code for the WACV 2022 paper,

"Hole-robust Wireframe Detection" by Naejin Kong, Kiwoong Park and Harshith Goka

Paper with Supplementary Materials

See arXiv Version.

Dataset Generation Scripts

See dataset/.

Algorithm Code

See algorithm/.

Tested Environment

• Ubuntu 18.04
• Python 3.7
• Virtualenv
• Nvidia GPU + Cuda 10.1

Installation

1. Clone this repository.
2. Set up virtualenv.

# Create a new virtualenv
$ sudo apt-get install python3.7-dev
$ virtualenv venv --python=python3.7

# Activate virtualenv
$ source venv/bin/activate

# Install packages
(venv) $ pip install --upgrade pip
(venv) $ pip install -r requirements.txt \
-f https://download.pytorch.org/whl/torch_stable.html \
-f https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html
(venv) $ pip uninstall pycocotools
(venv) $ pip install pycocotools==2.0.2 --no-binary pycocotools

License

Please refer to LICENSE.

Citing

@InProceedings{Kong_2022_WACV,
    author    = {Kong, Naejin and Park, Kiwoong and Goka, Harshith},
    title     = {Hole-Robust Wireframe Detection},
    booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)},
    month     = {January},
    year      = {2022},
    pages     = {1636-1645}
}