Generative Large Model-Driven Methodology for Color Matching and Shape Design in IP Products

Area of Science:

• Design Science
• Artificial Intelligence
• Sustainable Tourism

Background:

• Generative large models and AI-generated content (AIGC) are transforming product design.
• Tourism intellectual property (IP) cultural products are vital for sustainable tourism.
• Existing design methodologies lack integration of generative large models for tourism IP products.

Purpose of the Study:

• To propose a practical methodology for color matching and shape design of tourism IP cultural products.
• To leverage multimodal generative large models in the design process.
• To enhance the appeal and effectiveness of tourism cultural products.

Main Methods:

• Utilized GPT-4o for exploring visitor emotional needs and identifying target imagery.
• Employed Midjourney for shape generation and quadratic curvature entropy for shape selection.
• Used Midjourney for color image generation, AHP and OpenCV for color selection, and color harmony calculations for combination.
• Conducted quantitative and qualitative evaluations using aesthetic measurement formulas and sensibility questionnaires.

Main Results:

• A novel methodology integrating multimodal generative large models for tourism IP product design was developed.
• The process effectively identified visitor emotional needs, generated suitable shapes, and selected optimal color combinations.
• Case study on harbor seal IP products demonstrated strong correlation between quantitative and qualitative evaluations.

Conclusions:

• The proposed methodology is effective for the color matching and shape design of tourism IP cultural products.
• This approach enhances sustainable tourism development through innovative product design.
• The integration of AI in cultural product design offers significant potential for the tourism industry.