SelfSupervisedFiseyeRectification

An official pytorch implementation of the paper "Self-Supervised Fisheye Image Rectification by Reconstructing Coordinate Relations"

Our network is based on single parameter division model, architecture is shown below.

Quick start

1. Clone the project.

   git clone https://github.com/MasakiHosono/SelfSupervisedFisheyeRectification.git
   

2. Install the dependencies.

   pip3 install -r requirements.txt
   

3. Prepare dataset Download cityscapes dataset and put it under data/ in the project root. Then run the following script.

   python3 tools/prepare_cityscapes.py
   

4. Training.

   python3 src/train.py cfg/cityscapes.yml
   

5. Testing.

   python3 src/test.py cfg/cityscapes.yml
   

Citation

@inproceedings{hosono2021self,
  title={Self-Supervised Deep Fisheye Image Rectification Approach using Coordinate Relations},
  author={Hosono, Masaki and Simo-Serra, Edgar and Sonoda, Tomonari},
  booktitle={2021 17th International Conference on Machine Vision and Applications (MVA)},
  pages={1--5},
  year={2021},
  organization={IEEE}
}