LF-EASR: Efficient Light Field Angular Super-Resolution With Sub-Aperture Feature Learning and Macro-Pixel Upsampling

This repository contains official pytorch implementation of Efficient Light Field Angular Super-Resolution With Sub-Aperture Feature Learning and Macro-Pixel Upsampling in TMM 2022, by Gaosheng Liu, Huanjing Yue, Jiamin Wu, and Jingyu Yang. TMM 2022 LF-EASR

Requirement

• Ubuntu 18.04
• Python 3.6
• Pyorch 1.7
• Matlab

Dataset

We provide MATLAB code for preparing the training and test data. Please first download light field datasets, and put them into corresponding folders. The real-world training data is available in SIGGRAPH/ACM Trans. Graph..

Training

• Run:

  python train.py

Test

We provide the pre-trained model for real-world scenes in ./checkpoint_SIG.

• Run:

  python test.py

Citation

If you find this work helpful, please consider citing the following papers:

@article{liu2022efficient,
  title={Efficient Light Field Angular Super-Resolution With Sub-Aperture Feature Learning and Macro-Pixel Upsampling},
  author={Liu, Gaosheng and Yue, Huanjing and Wu, Jiamin and Yang, Jingyu},
  journal={IEEE Transactions on Multimedia},
  year={2022},
  publisher={IEEE}
}

Acknowledgement

Our work and implementations are based on the following projects:
LF-DFnet
LF-InterNet
We sincerely thank the authors for sharing their code and amazing research work!