This study addresses the challenges of precise color formulation in 3D food printing. Traditional trial-and-error methods for color matching are time-consuming and resource-intensive. To overcome this, we developed a data-driven expert system to identify an effective method for ingredient recommendation. Using an experimental database, including precise ingredient proportions (e.g., base materials and colorants) and their corresponding measured color values (RGB, CIELAB, CMYK, HSV), we evaluate the system’s accuracy in retrieving suitable printing formulas. Our findings inform a decision-support graphical user interface (GUI) tool that assists food technologists in making informed formulation choices. Experimental results demonstrate that the Euclidean distance algorithm achieves a strong balance in matching precision, successfully retrieving formulas within an acceptable human visual threshold (Delta E <= 3.0). Furthermore, multiple linear regression analysis reveals a highly predictable relationship between colorant concentrations and chromaticity values (Lab*). Although the current database-lookup approach yields highly accurate results for predefined formulas, its inability to extrapolate limits its deployment for entirely unseen colors. Consequently, this study establishes a robust data pipeline, and we recommend integrating this system with predictive Machine Learning models in the future. This approach offers a compelling trade-off between current computational reliability and future generative capabilities, significantly advancing automated color matching in the alternative protein and food technology sectors.

Punnarust Silparattanawong