Color Prediction Techniques in Textile Engineering

Accurate prediction of textile colour is fundamental to sustainable and efficient production, reducing waste of water, energy and dyestuffs. Colour prediction algorithms bridge the gap between chemical processes, fibre composition and perceptual output by modelling interactions of light with dyed or blended fibres. Traditional approaches rely on theoretical frameworks such as the Kubelka–Munk model to relate reflectance spectra to dye concentration, while more recent methods employ intelligent algorithms and physics-informed machine learning to capture complex, non-linear dyeing dynamics. Modern developments integrate optimisation techniques, deep neural architectures and advanced colour difference metrics to enhance prediction accuracy across diverse substrates, from uniform melange yarns to bicomponent filaments. The convergence of data-driven and physically constrained methods promises more reliable recipe formulation, faster prototyping and improved consistency for global textile supply chains.

Research from Nature Portfolio

Recent studies have compared foundational reflectance models, demonstrating that the Kubelka–Munk approach retains superior simplicity and reliability when estimating dye concentration and spectral reflectance across common textile substrates, even when alternative formulations are extended. A novel application of genetic algorithms has optimised colour matching for bicomponent (PET/PTT) filaments, yielding dye recipes that minimise perceptual colour difference with high precision and reduced computational burden. In denim fabrication, an ant colony optimisation framework has been developed to harmonise reactive and disperse dye processes on cotton–polyester blends, achieving uniform shades by predicting optimal dyestuff combinations that satisfy both mechanical and chromatic requirements under varying dyeing protocols.

Color Prediction Techniques in Textile Engineering publication trend

The graph below shows the total number of articles in color prediction techniques in textile engineering across all publications each year (not limited to Nature Index journals).

Technical terms

Kubelka–Munk model: A two-flux theory relating diffuse reflectance of a dyed substrate to its absorption and scattering coefficients for predicting colour outcomes.

Allen–Goldfinger model: A variation of the Kubelka–Munk framework that incorporates Beer–Lambert absorption coefficients for spectral reflectance estimation.

CIELAB colour space: A three-dimensional perceptually uniform colour coordinate system (L*, a*, b*) widely used for quantifying and comparing textile colours.

Genetic algorithm: An optimisation technique inspired by natural selection that iteratively evolves dye recipe parameters to minimise colour difference metrics.

Ant colony algorithm: A bio-inspired optimisation method based on the foraging behaviour of ants, used to predict optimal dye combinations in multi-phase processes.

Physics-informed modelling: A machine learning approach that embeds physical laws or empirical colour difference formulas into the training process to improve predictive accuracy.

Transformer network: A deep learning architecture employing attention mechanisms to model complex interactions, here applied to spectral and colour sequence prediction.

CMC colour difference formula: A metric developed by the Colour Measurement Committee for quantifying perceptual differences between two colours under controlled viewing conditions.

References

1. Estimation of dye concentration by using Kubelka–Munk and Allen–Goldfinger reflective models: comparing the performance. Scientific Reports (2023).
2. Color matching of bicomponent (PET/PTT) filaments with high performances using genetic algorithm. Scientific Reports (2024).
3. Dyeing of advanced denim fabrics (blend of cotton/bicomponent polyester filaments) using different processes and artificial intelligence method. Scientific Reports (2024).
4. Leveraging multi-output modelling for CIELAB using colour difference formula towards sustainable textile dyeing. Autonomous Intelligent Systems (2024).
5. Intelligent techniques and optimization algorithms in textile colour management: a systematic review of applications and prediction accuracy. Fashion and Textiles (2024).
6. Spectral prediction method based on the transformer neural network for high-fidelity color reproduction.. Optics Express (2024).