Book Section #65098

Fang, Fufu, Gong, Han, Mackiewicz, Michal and Finlayson, Graham (2017) UNSPECIFIED In: UNSPECIFIED Korea Society of Color Studies, Jeju, Korea. ISBN 978-89-5708-276-8

[img]
Preview
PDF (Accepted manuscript) - Submitted Version
Available under License ["licenses_description_unspecified" not defined].

Download (587kB) | Preview

    Abstract

    For a camera image, the RGB response from the imaging sensor cannot be used to drive display devices directly. The reason behind this is two-fold: different cameras have different spectral sensitivities, and there are different target output spaces (e.g. sRGB, Adobe RGB, and XYZ). The process of mapping from captured RGBs to an output colour space is called colour correction. Colour Correction is of interest in its own right (e.g. for colour measurement), but it is also an important part of the colour processing pipelines found in digital cameras. In this paper, we look at the problem of mapping device RGB values to corresponding CIE XYZ tristimuli. We make three contributions. First, we review and implement a range of colour correction algorithms. We benchmark these algorithms in experiments using both synthetic data (so we can numerically assess a wider range of cameras) and real image data. In our second contribution, we develop an ensemble method to combine colour correction algorithms to further enhance performance. For the methods tested, we find there is small extra power in combining the methods. Our final — and perhaps most important contribution — is to provide an open source colour correction MATLAB toolbox for the community, implementing the algorithms described in the paper. As well, all our experimental data is provided.

    Item Type: Book Section
    Uncontrolled Keywords: colour correction ,toolbox
    Faculty \ School: Faculty of Science
    Faculty of Science > School of Computing Sciences
    Related URLs:
    Depositing User: Pure Connector
    Date Deposited: 06 Oct 2017 08:14
    Last Modified: 16 Feb 2019 01:07
    URI: https://ueaeprints.uea.ac.uk/id/eprint/65098
    DOI:

    Actions (login required)

    View Item