AUTHORS: Oleg Milder, Dmitry Tarasov, Andrey Tyagunov

Download as PDF

ABSTRACT: Novel methods of digital image processing demand various approaches to spectral reflection prediction. The continuing complexity of the models leads to high requirements for computing power; however, this does not always contribute to the convenience and accuracy of forecasts. We offer an easy-to-use method for solving the direct problem of spectral reflection prediction using artificial neural networks. For color practitioners, prediction accuracy in terms of color difference is highly important. For researchers of artificial intelligence, the organization of the network learning process is of overriding interest. Those and other interoperates the necessary and sufficient minimum of the training sample to ensure satisfactory forecast accuracy. In this paper, we determine the size of such a sample. When training a network, we use spectral density instead of spectra. This provides a simplified simulation and improved accuracy of the forecast, which is confirmed experimentally.

KEYWORDS: Spectrum, Color, Prediction, Artificial neural network, Image processing

REFERENCES:

[1] Bala R. (2003) Device characterization. Digital Color Imaging Handbook, ed. G. Sharma (CRC Press, Boca Raton, FL), 269–379.

[2] Kubelka P. & Munk F. (1931) Ein Beitrag zur Optik der Farbanstriche, Zeitschrift für technische Physik, Germany, 12, 593–601.

[3] Neugebauer H.E.J. (1937) Die theoretischen Grundlagen des Mehrfarbendrucks. Zeinschrift fur Wissenschaftliche Photographie Photophysik Photochemie, 36, 36–73.

[4] Yule J. A. C. & Nielsen W. J. (1951) The penetration of light into paper and its effect on halftone reproductions. Proc. TAGA Conference 1951 (TAGA, Sewickley, PA), 65– 76.

[5] Viggiano J. A. S (1990) Modeling the color of multi-colored halftones. Proc. TAGA Conference 1990 (TAGA, Sewickley, PA), 44– 62.

[6] Hersch R.D. & Crété F. (2005) Improving the Yule–Nielsen modified spectral Neugebauer model by dot surface coverages depending on the ink superposition conditions. Proc. SPIE 5667, 434–445.

[7] Mazauric S., Hebert M., & Fournel T. (2018) Revisited Yule–Nielsen model without fitting of the n parameter. Journal of the Optical Society of America, 35, no.2, 244–255.

[8] Wyble D.R. & Berns R.S. (2000) A critical review of spectral models applied to binary color printing. Color Research & Application, 25, 4–19.

[9] Pauli H. (1976) Proposed extension of the CIE recommendation on 'Uniform color spaces, color difference equations, and metric color terms'. Journal of the Optical Society of America, Vol. 66, 866–867

[10] Hunt, R.W.G. (2004). The Reproduction of Colour, 6th rev. ed. Wiley & Sons, NJ. 724p.

[11] Hill B., Roger Th., Vorragen F.W. (1997) Comparative Analysis of the Quantization of Color Spaces on the Basis of the CIELAB Color-Difference Formula. ACM Transactions on Graphics, 16(2), 109–154.

[12] Milder O.B., Tarasov D.A., Titova M.Yu (2017) Inkjet Printers Linearization Using 3D Gradation Curves. CEUR Workshop Proceedings. Proceedings of the 1st International Workshop on Radio Electronics & Information Technologies (REIT 2017), Yekaterinburg, Russia, March 15, 2017. Vol.1814. 74–83.

[13] Milder O., Tarasov D. (2018) Ink-jet printer’s characterization by 3D gradation trajectories on an equidistant color difference basis. Lecture Notes in Computer Science (LNCS). Vol. 10749. 40–52.

[14] Milder O., Tarasov D. (2018) Gradation surfaces as a method for multi-color ink-jet printers color specifications management. Lecture Notes in Computer Science (LNCS). Vol. 10749. 53–61.

[15] Kayri M. (2016) Predictive abilities of Bayesian regularization and levenbergmarquardt algorithms in artificial neural networks: A comparative empirical study on social data. Mathematical and Computational Applications. 21(2), 21020020

[16] Hersch R.D., Hebert M. (2015). Base models for color halftone reproduction. M. Kriss. Handbook of Digital Imaging, 2, Wiley, 1079– 1132.