This repository provides an enhanced and unified toolbox for Light Field (LF) Image Super-Resolution (SR), supporting both:
- LFSSR – LF Spatial SR (improving resolution of each sub-view image)
- LFASR – LF Angular SR (increasing angular resolution and enabling novel view synthesis)
This is an extension of our previous work, BasicLFSR, which focused solely on spatial SR. It has proven to be a helpful toolbox for researchers to quickly get started with LF spatial SR and to facilitate the development of new algorithms.
Looking ahead, BasicLFSR-plus, together with our BasicLFDisp repository for LF disparity estimation, aims to provide a more comprehensive and user-friendly benchmark toolbox for the LF research community.
git clone https://github.com/ZhengyuLeung/BasicLFSR-plus.git
cd BasicLFSR-plus
pip install -r requirements.txt💡 Tip: Make sure to use a Python virtual environment (e.g., conda or venv) to avoid package conflicts.
To get started with a specific task, please refer to the corresponding README:
| Task | Description | Instructions |
|---|---|---|
lfssr |
Code, models and pretrained weights for LF spatial SR | See README_LFSSR |
lfasr |
Code, models and pretrained weights for LF angular SR | See README_LFASR |
Feel free to open pull requests or discussions!
Welcome to raise issues or email to zyliang@nudt.edu.cn for any question regarding our BasicLFSR-plus.