INTEL-TUT Dataset for Camera Invariant Color Constancy Research

Notifications

1) This dataset is intended only for academic use. 2) The image contents must not be reproduced or made public in any other form than via academic publications and conferences.

Citation

Any work that uses the dataset should cite the following publication:

https://arxiv.org/abs/1703.09778

Introduction

INTEL-TUT dataset is designed for camera invariant color constancy research. Camera invariance corresponds to the robustness of an algorithm’s performance when run on images of the same scene taken by different cameras. Accordingly, images in the database correspond to several lab and field scenes each of which are captured by three different cameras with minimal registration errors. The lab scenes are also captured under five different illuminations. The spectral responses of cameras and the spectral power distributions of the lab light sources are also provided, as they may prove beneficial for training future algorithms to achieve color constancy. As a side contribution, this dataset also includes images taken by a mobile camera with color shading corrected and uncorrected results. This allows research on the effect of color shading as well.

Organization of the Dataset

The images in the dataset are organized as follows:

'RootPath/CameraName/SetName'

Note that, for the Mobile camera only, there exists a color shading correction (CSC) option.

'RootPath/CameraName/SetName/CSCinfo'

CameraName: Indicates the name of the camera used to take the image, and it is one of the following: 1) Canon_5DSR 2) Nikon_D180 3) Mobile

SetName: Indicates the name of the set (category) of images and can take following names: 1) field 2) field2 (Valid for only Canon_5DSR) 3) lab_printouts 4) lab_realscene

CSCinfo: Indicates if color shading correction is applied on the images or not and is only valid for Mobile camera. This is indicated as follows: 1) With_BLC_CSC 2) Without_BLC_CSC

Contents and Annotation

Within each folder, the contents are named as follows:

'SetName_SceneNumber_LightSourceName_CameraName'

The contents are as follows for each image, and follow the above annotation with corresponding extensions:

1)RAW Image: These images underwent no processing. Extension: '_RAW.tiff'

2)High Resolution RGB: These images underwent demosaicing, black level correction, saturation point stretching operations. Extension: '_RGB.tiff'

3)1080p RGB: These images underwent demosaicing, black level correction, saturation point stretching operations, bilinear downscaling to 1080p. Extension: '_RGB_1080.tiff'

4)White Points: This file contains the ground truth illumination chromaticity in l2 normalized RGB format. Extension: '_wp'

5)Color Correction Matrix: This file contains the color correction matrix from sensor RGB to sRGB. It is stored as a 1x9 vector. Extension: '_ccm'

6)JPEG Image: These images underwent demosaicing, black level correction, saturation point stretching operations, white balancing, color correction, 0.45 gamma correction and minor sharpening and brightness normalization. These images are only shared for illustration of how a color corrected image should look like. Extension: '.jpg'

Supplementary Material

Supplementary material can be found in the 'Supplementary' folder and contains the following:

1) Lab Light Source Data: Contains lux, CCT, CIEx, CIEy values of light sources measured at several time instances after switching on the light sources. Also contains color correction matrices for the light sources for each camera.

In the file 'Lab_light_source_data.xlsx'

2) Spectral Light Source Data: Contains the spectral distribution of each light source at 360-780 nm interval with 1 nm steps.

In the file 'Spectral_light_source_data.xlsx'

3) Spectral Responses of the Cameras: Contains the spectral response of the cameras at 380-780 nm interval with 5 nm steps.

In the file 'Spectral_responses.xlsx'

Citation and Credits

1) Any work that uses the dataset should cite the following publication:

https://arxiv.org/abs/1703.09778

2) Jarno Nikkanen, Vitali Samurov and Anttu Koski, from Intel (Tampere, Finland Site) for capturing the images.

3) For the decoding implementation of Canon and Nikon RAW files we have used: www.cybercom.net/~dcoffin/dcraw/