Inria synthetic light field datasets
(with disparity maps)

Contacts: J. Shi, X. Jiang, C. Guillemot

The datasets were generated in order to test depth estimation method in light field. They were submitted along with our proposed depth estimation framework in TIP, July 2019.

The 'Inria synthetic light field datasets' is composed of two sub-datasets (SLFD & DLFD) generated with open source software Blender. Sparse Light Field Dataset (SLFD) includes 53 scenes with disparity range [-20,20] pixels between adjacent views, and Dense Light Field Dataset (DLFD) contains 39 scenes with disparity range [-4,4] pixels. Both datasets have light fields of spatial resolution 512 x 512 and of angular resolution 9 x 9. Besides color image (format .png) of each view, corresponding disparity map is also available in format .npy (for Python users) or .mat (for Matlab users).

For each light field in dataset, the views are numbered as depicted in the diagram below, it should be noted that the number starts from 1, not 0.

Generated light field of dimension 9 x 9

The following videos show several light fields examples, each scene has a color map video and a disparity map video.

Sparse Light Field Dataset

Furniture [-13.55,12.61] Lion [-3.19,14.38] Toy_bricks [-0.40,10.94] Electro_devices [-4.85,8.23]

Dense Light Field Dataset

White_roses [-1.52,3.38] Bowl & Chair [-1.87,-0.04] Kitchen_board [-1.61,-0.10] Toy_friends [-2.02,2.54]

The datasets are available here : We will appriciate it if you can cite our paper when using the datasets.


This work has been supported by the EU H2020 Research and Innovation Programme under grant agreement No 694122 (ERC advanced grant CLIM).


Creative Commons LicenseAll datasets and benchmarks on this page are copyright by us and published under the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License. This means that you must attribute the work in the manner specified by the authors, you may not use this work for commercial purposes and if you alter, transform, or build upon this work, you may distribute the resulting work only under the same license.