Title: Applying Artificial Intelligence in Style Transfer to Recreate Assets in Vietnamese History Games
Authors: Khai Nghiem Tuan Duc Nguyen Minh
Volume: 9
Issue: 10
Pages: 144-151
Publication Date: 2025/10/28
Abstract:
The digital entertainment sector continues to grow, along with the demand for digital content that captures a slice of history or ancient culture. The development of computer games themed on history, however, faces a challenge: the lack of culturally sensitive, culturally integrated game assets, especially those that portray the goals of various periods of Vietnamese history. The gap undercuts the potential of culturally sensitive creative design and the risk of cultural misappropriation or distortion of the varied cultural legacy of Vietnam. To bridge this gap, the current work presents a proposed AI-based approach that combines a Generative AI model with style transfer methodologies to modify archival materials from other cultures, creating Vietnamese-inspired variants that reflect Vietnamese cultural aesthetics. The proposed approach uses Stable Diffusion, an image-generative AI, which the author has optimized for use on CPU-only machines and for low RAM GPUs. The use of the Gemini Large Language Model (LLM) is an additional novel aspect of this work, where Gemini engages in 'prompt engineering' or converting simple user intentions into rich, contextually appropriate prompts for Stable Diffusion. This significantly enhances the creative work and reduces the game designers' need for technical knowledge and skills. The system demonstrates the production of culturally relevant and visually coherent Vietnamese-style game assets and significant decreases in design time and effort. This research moves beyond a technical contribution by offering a comprehensive and practical approach to the preservation of Vietnamese digital heritage and the creation of culturally-informed bundled video game content. Additionally, this work presents a reproducible prototype and a procedural reference for cross-disciplinary Artificial Intelligence research, contemporary digital art, and cultural heritage preservation.