https://github.com/byronhsu/fegan

Self-Supervised Deep Learning for Fisheye Image Rectification (pytorch implementation).
https://github.com/byronhsu/fegan

computer-vision deep-learning fisheye gan image-rectification self-supervised-learning unsupervised-learning

Last synced: about 1 month ago
JSON representation

Self-Supervised Deep Learning for Fisheye Image Rectification (pytorch implementation).

• Host: GitHub
• URL: https://github.com/byronhsu/fegan
• Owner: ByronHsu
• Created: 2019-09-17T09:06:56.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2019-10-15T13:20:34.000Z (over 5 years ago)
• Last Synced: 2024-12-16T00:41:31.369Z (about 2 months ago)
• Topics: computer-vision, deep-learning, fisheye, gan, image-rectification, self-supervised-learning, unsupervised-learning
• Language: Python
• Homepage:
• Size: 5.93 MB
• Stars: 42
• Watchers: 3
• Forks: 5
• Open Issues: 4
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# Self-Supervised Deep Learning for Fisheye Image Rectification



This is the official pytorch implementation of [Self-Supervised Deep Learning for Fisheye Image Rectification](). The code is written by [Byron Hsu](https://github.com/ByronHsu) and [Brian Chao](https://github.com/Mckinsey666) from Department of Electrical Engineering, National Taiwan University. This work is currently in the submission to [2020 ICASSP](https://2020.ieeeicassp.org/). 

We welcome any advice for this work. Feel free to send issue or PR!

## Installation

1. Clone the project

```

git clone ...

```

2. Install the prerequisites

```

pip3 install -r requirements.txt

```

## Training

1. Start training!

```

bash train_gcgan.sh [dataset_path] [name]

```

2. Monitor the training process by tensorboard

```

tensorboard --logdir runs

```

## Testing

```

bash test_gcgan.sh [dataset_path] [name] [which_epoch]

```

## Reference

- [pytorch-CycleGAN-and-pix2pix](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix)

- [GCGAN](https://github.com/hufu6371/GcGAN).