Dataset for multispectral illumination estimation using deep unrolling network

Thumbnail
Name:
KAUST_SpectralReflectanceImages_h5.zip
Size:
11.49Gb
Format:
application/zip
Description:
Spectral reflectance image data in h5 format
Download
Type
Dataset
Authors
Li, Yuqi
Fu, Qiang cc
Heidrich, Wolfgang cc
KAUST Department
Visual Computing Center (VCC)
Date
2021-08-02
Permanent link to this record
http://hdl.handle.net/10754/670368

Metadata
Show full item record
Summary
Our dataset repository includes 409 spectral reflectance images. The images consist of various indoor and outdoor scenes acquired with a compact scanning-based hyperspectral camera: Specim IQ. The indoor scenes include clothes, papers, toys, vegetables, fruits, optical elements, etc. and the outdoor scenes include buildings, plants, vehicles, animals, etc. The captured images have a spatial resolution of 512×512 pixels and 34 spectral bands ranging from 400nm to 730nm.

Details of the database can be found in the following publication:

@article{Yuqi2021SpecSeperation,
title={Multispectral illumination estimation using deep unrolling network},
author={Li, Yuqi and Fu, Qiang and Heidrich, Wolfgang},
booktitle={2021 IEEE International Conference on Computer Vision(ICCV)},
pages={1--8},
year={2021},
organization={IEEE} }

We capture various scenes containing a whiteboard with flat spectral reflectance and calculate the reflectance spectral images. The spectral reflectance images are stored as H5 files.

Read the H5 data file (Matlab example):

data = h5read('*.h5','/img\');
Citation
Li, Y., Fu, Q., & Heidrich, W. (2021). Dataset for multispectral illumination estimation using deep unrolling network [Data set]. KAUST Research Repository. https://doi.org/10.25781/KAUST-6930V
Sponsors
This work was supported by the KAUST baseline funding.
Publisher
KAUST Research Repository
DOI
10.25781/KAUST-6930V
Relations
Is Supplement To:
Collections
Datasets

The following license files are associated with this item:

Attribution-NonCommercial 3.0 United States
Except where otherwise noted, this item's license is described as Attribution-NonCommercial 3.0 United States