An AIGC-empowered methodology to product color matching design

Abstract

With the emergence of various generative AI applications, artificial intelligence-generated content (AIGC) demonstrates positive potential for design activities. However, few scholars have proposed a practical AIGC-based design methodology. This paper introduces an AIGC-empowered methodology for product color-matching design. ChatGPT generates target imageries describing the design features, and Midjourney constructs a shape sample database based on these imageries, identifying representative shapes through consumer perceptual questionnaires. Meanwhile, three Midjourney-based color image generation methods are proposed to generate color images that match the target imageries, and an application is developed to extract the dominant colors from these images. Finally, based on the color harmony theory, the optimal color combinations are derived from the extracted dominant colors and applied to the representative shape to generate color-matching alternatives. AHP-based expert evaluation (Evaluation_1) and consumer perceptual evaluation (Evaluation_2) are employed to ascertain the optimal design solution. This paper uses the household vacuum cleaner as an example to demonstrate the proposed color-matching methodology. Consistent evaluation results are obtained, supported by a significant Pearson correlation coefficient. Research findings suggest that AIGC has the potential to revolutionize traditional product color-matching design. Additionally, the methodology highlights the collaborative potential between generative AIs and conventional computer-aided design tools.