In this thesis, we proposed a straightforward data-driven approach that can be easily adopted. The method can predict the concentrations of a chosen dye combination and reproduce the color of the target fabric.
The approach leverages a color prediction model implemented by deep neural networks. We find the inverse value, which is the recipe, of the model with the CIEL∗a∗b∗ or reflectance of a given fabric. And we use a soft proofing model to predict the color, that is, CIEL∗a∗b∗ or reflectance of the predicted recipe.
We use two different ways to find the inverse values of the color prediction models, grid search and gradient descent. Grid search performs better in finding the inverse value but both perform almost equally well in ∆Eab∗ .
Our approach is flexible in the size of the recipes, which is between 1 to 4 and it can be easily scaled up. The way we use in the approach can avoid metamerism problems that will show up in some methods. Moreover, our approach can still have acceptable performances on novel recipes, which are not been discussed in other researches.
Under the structure of our approach, possible improvements that can be done are perhaps mainly in feature engineering. Currently, fabrics are presented as one-hot vectors in the recipe vectors. One can study the physical properties of fabrics
and might be able to extract more features that might be helpful. A more balanced training set might also be helpful. Adding anchor points to each dye combinations might be helpful too since it may decrease the possibility of extrapolation.
The approach doesn’t be proposed to replace the colorists and there is no way that the approach can replace colorists. Instead, it is built to assist the colorists to find recipes faster. The approach proposed in this thesis is already adopted by Everest Textile, and we hope it can be helpful to others in the industry, too.
Reference
[1] Paul Kubelka and Franz Munk. “An article on optics of paint layers”. In: Z.
Tech. Phys 12.593-601 (1931).
[2] David L MacAdam. “Visual sensitivities to color differences in daylight”. In:
Josa 32.5 (1942), pp. 247–274.
[3] JL Saunderson. “Calculation of the color of pigmented plastics”. In: JOSA 32.12 (1942), pp. 727–736.
[4] Paul Kubelka. “New contributions to the optics of intensely light-scattering materials. Part I”. In: Josa 38.5 (1948), pp. 448–457.
[5] DG Nickols and SE Orchard. “Precision of Determination of Kubelka and Munk Coefficients from Opaque Colorant Mixtures”. In: JOSA 55.2 (1965), pp. 162–164.
[6] Eugene Allen. Colorant Formulation and Shading in Optical Radiation Mea-surement Vol. 2 Color MeaMea-surement, Gram F and Bartleson CJ. 1980.
[7] Leo Breiman, Jerome Friedman, Charles J Stone, et al. Classification and regression trees. CRC Press, 1984.
[8] JM Bishop, MJ Bushnell, and S Westland. “Application of neural networks to computer recipe prediction”. In: Color Research & Application 16.1 (1991), pp. 3–9.
[9] Shoji Tominaga. “A neural network approach to color reproduction in color printers”. In: Color and Imaging Conference. Vol. 1993. 1. Society for Imaging Science and Technology. 1993, pp. 173–177.
[10] Andrew S Glassner. Principles of digital image synthesis: Vol. 1. Vol. 1. Else-vier, 1995.
[11] Shoji Tominaga. “Color control of printers by neural networks”. In: Journal of Electronic Imaging 7.3 (1998), pp. 664–672.
[12] Stephen Westland. “Artificial neural networks and colour recipe prediction”.
In: Proceedings of the International Conference and Exhibition: Colour Science.
1998, pp. 225–233.
[13] Andy Liaw, Matthew Wiener, et al. “Classification and regression by random-Forest”. In: R news 2.3 (2002), pp. 18–22.
[14] János Schanda. Colorimetry: Understanding the CIE System. John Wiley &
Sons, 2007.
[15] A Shams Nateri and Ehsan Ekrami. “Dye binary mixture formulation by means of derivative ratio spectra of the Kubelka–Munk function”. In: Color Research & Application: Endorsed by Inter-Society Color Council, The Colour Group (Great Britain), Canadian Society for Color, Color Science Association of Japan, Dutch Society for the Study of Color, The Swedish Colour Centre Foundation, Colour Society of Australia, Centre Français de la Couleur 35.3 (2010), pp. 193–199.
[16] Hongying Yang, Sukang Zhu, and Ning Pan. “On the Kubelka—Munk Single-Constant/Two-Constant Theories”. In: Textile Research Journal 80.3 (2010), pp. 263–270.
[17] J Militkỳ. “Fundamentals of soft models in textiles”. In: Soft Computing in Textile Engineering. Elsevier, 2011, pp. 45–102.
[18] Elham Sadat Yazdi Almodarresi, Javad Mokhtari, Seyed Mohammad Taghi Almodarresi, et al. “A scanner based neural network technique for color match-ing of dyed cotton with reactive dye”. In: Fibers and polymers 14.7 (2013), pp. 1196–1202.
[19] Bahar Sennaroglu, Erhan Öner, and Ö Şenvar. “Colour recipe prediction in dyeing acrylic fabrics with fluorescent dyes using artificial neural network”.
In: Industria textilă 65 (Jan. 2014), pp. 22–28.
[20] Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. “Deep learning”. In: nature 521.7553 (2015), pp. 436–444.
[21] Alexander Mordvintsev, Christopher Olah, and Mike Tyka. Inceptionism: Go-ing Deeper into Neural Networks. 2015. url: https://research.googleblog.
com/2015/06/inceptionism-going-deeper-into-neural.html.
[22] Michael A Nielsen. Neural networks and deep learning. Vol. 2018. Determina-tion press San Francisco, CA, 2015.
[23] Jürgen Schmidhuber. “Deep learning in neural networks: An overview”. In:
Neural networks 61 (2015), pp. 85–117.
[24] Tianqi Chen and Carlos Guestrin. “XGBoost: A Scalable Tree Boosting Sys-tem”. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. KDD ’16. San Francisco, California, USA: Association for Computing Machinery, 2016, pp. 785–794.
[25] Mei-Yun Chen, Ci-Syuan Yang, and Ming Ouhyoung. “A Smart Palette for Helping Novice Painters to Mix Physical Watercolor Pigments.” In: Eurograph-ics (Posters). 2018, pp. 1–2.
[26] Ya-Bo Huang, Mei-Yun Chen, and Ming Ouhyoung. “Perceptual-based CNN model for watercolor mixing prediction”. In: ACM SIGGRAPH 2018 Posters.
2018, pp. 1–2.
[27] Oleg B Milder and Dmitry A Tarasov. “Spectral Reflection Prediction by Ar-tificial Neural Network”. In: CEUR Workshop. Proceedings of the 3rd Inter-national Workshop on Radio Electronics & Information Technologies, Yeka-terinburg, Russia. Vol. 2076. 2018, pp. 86–95.
[28] Dmitry Tarasov, Oleg Milder, and Andrey Tyagunov. “Inverse problem of spectral reflection prediction by artificial neural networks: Preliminary re-sults”. In: 2018 Ural Symposium on Biomedical Engineering, Radioelectronics and Information Technology (USBEREIT). IEEE. 2018, pp. 144–147.
[29] Steven HH Ding, Benjamin CM Fung, and Philippe Charland. “Asm2vec:
Boosting static representation robustness for binary clone search against code obfuscation and compiler optimization”. In: 2019 IEEE Symposium on Secu-rity and Privacy (SP). IEEE. 2019, pp. 472–489.
[30] Dmitry Tarasov and Oleg Milder. “The Inverse Problem of Spectral Reflection Prediction by Artificial Neural Networks: Neugebauers Primaries vs. Recipes”.
In: 2019 International Multi-Conference on Engineering, Computer and Infor-mation Sciences (SIBIRCON). IEEE. 2019, pp. 0580–0583.