Alessandro Artusi, Alexander WilkieORCID iD
Color Printer Characterization Using Radial Basis Function Networks
In Proceedings Colour Imaging Conference: Device-Independent Colour, Colour Hardcopy, and Graphics Arts VI, IST&SPIE, Electronic Imaging, pages 70-80. January 2001.
[paper]

Information

  • Publication Type: Conference Paper
  • Workgroup(s)/Project(s):
  • Date: January 2001
  • ISBN: 0-8194-3978-9
  • Publisher: SPIE
  • Note: SPIE Conference, San Jose California, January 2001
  • Location: San Jose (USA)
  • Booktitle: Proceedings Colour Imaging Conference: Device-Independent Colour, Colour Hardcopy, and Graphics Arts VI, IST&SPIE, Electronic Imaging
  • Pages: 70 – 80
  • Keywords: neural network, Colorimetric characterization

Abstract

Colorimetric characterization is one step in the colorimetric reproduction process that permits faithful image reproduction across different devices. Its goal is to define a mapping function between the device--dependent color spaces in question (such as RGB or CMYK) and device--independent colour spaces (such as CIELAB or CIEXYZ), and vice versa. The work presented in this paper is an application study of utilizing radial basis function networks for the problem of colorimetric characterization of printer devices. The work we present is novel in seven ways: to begin with, this is the first work that uses radial basis function networks to resolve the colorimetric characterization of printers. Second, we used a new learning model to train such networks; our approach is based on a proposal by Carozza. Third, we use only 125 measured samples for the training of the network. Fourth, the computational costs for this training are very low when compared to previous techniques and allow to use this model in consumer products. Fifth, it is a general model which one can also use to define other transformations between color spaces. Sixth, it is possible to have a fast recharacterization of the device because the computational cost of the training phase and the number of training samples are low. Finally, it improves on the performance of multiple polynomials regression and tetrahedral interpolation.

Additional Files and Images

Weblinks

No further information available.

BibTeX

@inproceedings{Artusi-2001-Col,
  title =      "Color Printer Characterization Using Radial Basis Function
               Networks",
  author =     "Alessandro Artusi and Alexander Wilkie",
  year =       "2001",
  abstract =   "Colorimetric characterization is one step in the
               colorimetric reproduction process that permits faithful
               image reproduction across different devices. Its goal is to
               define a mapping function between the device--dependent
               color spaces in question (such as RGB or CMYK) and
               device--independent colour spaces (such as CIELAB or
               CIEXYZ), and vice versa. The work presented in this paper is
               an application study of utilizing radial basis function
               networks for the problem of colorimetric characterization of
               printer devices. The work we present is novel in seven ways:
               to begin with, this is the first work that uses radial basis
               function networks to resolve the colorimetric
               characterization of printers. Second, we used a new learning
               model to train such networks; our approach is based on a
               proposal by Carozza. Third, we use only 125 measured samples
               for the training of the network. Fourth, the computational
               costs for this training are very low when compared to
               previous techniques and allow to use this model in consumer
               products. Fifth, it is a general model which one can also
               use to define other transformations between color spaces.
               Sixth, it is possible to have a fast recharacterization of
               the device because the computational cost of the training
               phase and the number of training samples are low. Finally,
               it improves on the performance of multiple polynomials
               regression and tetrahedral interpolation. ",
  month =      jan,
  isbn =       "0-8194-3978-9",
  publisher =  "SPIE",
  note =       "SPIE Conference, San Jose California, January 2001",
  location =   "San Jose (USA)",
  booktitle =  "Proceedings Colour Imaging Conference: Device-Independent
               Colour, Colour Hardcopy, and Graphics Arts VI, IST&SPIE,
               Electronic Imaging",
  pages =      "70--80",
  keywords =   "neural network, Colorimetric characterization",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2001/Artusi-2001-Col/",
}