Customizable pattern synthesis: a deep generative approach for lantern designs.

Area of Science:

• Computer Vision
• Artificial Intelligence
• Cultural Heritage Preservation

Background:

• Pattern design is crucial for traditional lantern production, embedding cultural and artistic significance.
• Existing methods for pattern generation may lack flexibility and the ability to integrate traditional aesthetics with modern design.
• Generative models offer potential for creating novel and customizable patterns.

Purpose of the Study:

• To develop an innovative generative model for customizable lantern patterns.
• To integrate classical aesthetics with modern design features using a generative adversarial network (GAN).
• To enhance design flexibility while preserving cultural authenticity in pattern creation.

Main Methods:

• Developed a generative adversarial network (GAN)-based framework for pattern generation.
• Trained the model on an extensive dataset of over 17,000 pattern images across ten categories.
• Employed noise vector hybridization and post-processing techniques for enhanced control and flexibility.

Main Results:

• Achieved a high Inception Score of 5.259, outperforming other GAN-based approaches.
• Demonstrated effective integration of traditional pattern elements with AI-driven design.
• The model provides enhanced design flexibility and accurate control over pattern production.

Conclusions:

• The developed GAN-based model successfully generates customizable lantern patterns.
• The approach effectively merges traditional artistic values with modern AI capabilities.
• This tool offers significant potential for modernizing lantern design while preserving cultural heritage.