This dataset is a synthetic wide-baseline light field (WLF) dataset. It is created by Université Libre de Bruxelles (mainly) 
and Zhejiang University of Technology. 

The WLF dataset consists of two parts: FlyingObjects part and Hand-designed part, which are stored in Hierarchical Data Format (HDF). 
Each hdf file is comprised of the light fields (9x9 angular image views) and the ground truth disparities (9x9 disparity maps). 

This dataset is originally created for training the deep learning-based models for depth estimation. You are also allowed to use 
this dataset for other tasks if possible. NOTE THAT this dataset is provided for the research purpose only. 
For any commercial purpose, please contact us.

If you use our dataset in your research, please cite our paper:

@inproceedings{Li2020,
author={Li, Yan and Wang, Qiong and Lafruit, Gauthier},
title={A Lightweight Depth Estimation Network for Wide-baseline Light Fields},
year={2020}
}