Magical Brush: A Symbol-Based Modern Chinese Painting System for Novices

Title of the Paper

Magical Brush: A Symbol-Based Modern Chinese Painting System for Novices

Paper Information

• Subject Area: Human-Computer Interaction, AI Applications, Digital Art Creation Systems
• Keywords: Digital Painting, Cultural Priors, Artificial Intelligence, Human-Computer Collaboration, Chinese Painting, Generative Models, Symbolization

Research Background and Problem

• Identified Problems or Challenges:

  1. While modern Chinese painting integrates traditional cultural elements with contemporary styles, its intricate lines and details make it challenging for beginners to learn.
  2. Existing AI painting tools rarely support modern Chinese painting generation and lack deep consideration of cultural elements, making it difficult for users to learn cultural knowledge during the creative process.
  3. Striking a balance between AI assistance and user participation, while enabling beginners to appreciate the cultural significance of modern Chinese painting, remains a critical challenge.

• Research Significance:

  • This study aims to develop a symbol-based modern Chinese painting system to help beginners better understand and create modern Chinese paintings while experiencing the cultural essence behind them.
  • The research not only enhances the user’s creative experience but also promotes the application of AI in cultural heritage preservation.

• Research Motivation and Related Work:

  • Modern generative AI models have performed well in image generation and design assistance but often fail to support domain-specific creation, such as modern Chinese painting.
  • While there has been progress in human-computer collaboration systems, gaps remain in beginner-friendliness and cultural heritage promotion.
  • By integrating cultural factors into generative AI models through symbolization, this paper proposes and develops a modern Chinese painting assistance system tailored for beginners.

Solution

• Method or Solution:

  • This paper designs a symbol-based modern Chinese painting system named "Magical Brush," which combines deep learning generative models with cultural symbols to assist users in creating modern Chinese paintings while learning relevant artistic and cultural knowledge.

• Innovations:

  1. Proposes a symbolized painting mode that integrates cultural elements from traditional Chinese symbols with AI generative models.
  2. Develops an interactive symbol brush feature supporting symbol generation, color adjustment, style replacement, and positioning.
  3. Provides a painting guidance function to help users understand the cultural meanings of symbols and the significance of using specific colors.

• Implementation Steps and Key Technologies:

  1. Collected and analyzed 2,625 modern Chinese paintings to extract and quantify the frequency of symbol usage, ultimately selecting 17 representative symbols.
  2. Utilized StyleGAN2 to generate symbol images and improved generation quality through differential enhancement.
  3. Developed a symbol brush, color adjustment interface, style replacement interface, and basic painting tools to form a comprehensive creation system.
  4. Integrated a painting guidance interface to recommend relevant symbols and explain the cultural symbolism of different colors.

Research Outcomes

• Specific Results:

  • "Magical Brush" serves as an easy-to-use modern Chinese painting tool for beginners, enabling them to create high-quality works without requiring advanced painting skills.
  • The system’s educational features allow users to learn and understand the cultural essence of modern Chinese painting during the creative process.

• Advantages Over Existing Solutions:

  1. Compared to traditional non-AI tools (e.g., Procreate), Magical Brush simplifies the learning curve and enhances the educational and entertaining aspects of painting.
  2. Compared to purely AI-generated tools (e.g., Stable Diffusion), Magical Brush increases user engagement and control over the final output.

• Experimental or Evaluation Results:

  • User studies demonstrate that Magical Brush significantly outperforms comparison tools in usability, educational functionality, user satisfaction, and controllability.
  • Qualitative analysis highlights Magical Brush’s role in inspiring user creativity, fostering cultural learning, and improving creative efficiency.
  • Expert evaluations recognize its potential in beginner education and cultural dissemination but also suggest the need for further improvement in symbol diversity and the coherence of generated results.

• Limitations and Future Directions:

  • The training data for the symbolized generative model is limited, constraining the variety and styles of generated symbols.
  • While the system raises the creative baseline for beginners, it has limited appeal for professional artists.
  • Future work could focus on enhancing symbol coherence, providing real-time interactive feedback, and supporting animation generation to optimize user experience and cultural expression.