Adaptive Pixel Aggregation for Joint Spatial and Angular Super-Resolution of Light Field Images

This repository contains official pytorch implementation of Adaptive Pixel Aggregation for Joint Spatial and Angular Super-Resolution of Light Field Images, published in Information Fusion, 2023, by Gaosheng Liu, Huanjing Yue, Kun Li, and Jingyu Yang.

Dataset

We use the Standford Lytro, Kalantari, and HCInew as the training set. The test sets in mat format can be downloaded from BaiduPan, code IIRL.

Code

Dependencies

• Ubuntu 18.04
• Python 3.6
• Pyorch 1.3.1 + torchvision 0.4.2 + cuda 92

Citation

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

@article{liu2024adaptive,
  title={Adaptive pixel aggregation for joint spatial and angular super-resolution of light field images},
  author={Liu, Gaosheng and Yue, Huanjing and Li, Kun and Yang, Jingyu},
  journal={Information Fusion},
  volume={104},
  pages={102183},
  year={2024},
  publisher={Elsevier}
}