AI-Generated Art Experience in Kaohsiung Museum of Fine Arts
9/22/2025
Generative AI (GenAI) is reshaping how art museums engage audiences by transforming traditional viewing into immersive, participatory experiences. The Kaohsiung Museum of Fine Arts embraced this innovation in its latest exhibition. Through AI technology, visitors can generate personalized character portraits in the artist’s signature style using their mobile devices, allowing them to “enter” the artwork and experience art not just as spectators but as co-creators. The generated portraits are also displayed on a large projection screen. This integration of GenAI deepens audience engagement, personalizes the museum experience, and demonstrates how technology can expand the boundaries of artistic expression.
Challenges
Delivering this type of interactive experience requires:
- High-Performance Edge AI Inference System: to apply artistic styles to generate visitor photos in real time, avoiding long waits.
- Powerful Memory for Instant Creativity: to support high-resolution image generation and instant output.
- Quiet Operation: to maintain the immersive atmosphere of the museum.
- Stable and Reliable System Design: to run the AI photo booth seamlessly throughout the exhibition, serving thousands of visitors daily.
- Independent Placement: the chassis must be deployable outside a server room, allowing more flexible exhibition planning.
- Compact Size: the chassis should be as small as possible, making it easier to conceal within exhibition decorations.
Advantech Solutions
To enable this AI-Generated Art experience, the Advantech AIR-420—with its compact 28.6L chassis supporting dual NVIDIA RTX 5090 GPUs, designated by our ISV partner—is the ideal computing engine for AI photo booths, combining high performance with flexible installation. Its multi-GPU architecture enables real-time style transfer and generative AI without latency, while AMD Ryzen™ 7000/9000 CPUs (up to 12 cores) and DDR5 memory up to 192GB sustain heavy parallel processing. The AIR-420 generates each exhibition visitor’s picture in approximately 10–15 seconds, enabling efficient service for up to 1,500 visitors per day during peak periods. Its large VRAM and PCIe Gen5 storage ensure smooth inference pipelines even under high-traffic conditions. Equipped with 10GbE LAN, 2.5GbE LAN, and 20Gbps USB-C, the AIR-420 connects seamlessly to cameras, displays, and printers, supporting instant sharing and archival integration. Smart fan algorithms minimize noise to preserve the museum ambiance, while redundant power ensures reliable 24/7 operation. Unlike bulky rackmount servers, the AIR-420’s workstation-style form factor can be independently placed or discreetly integrated into exhibition setups, maintaining aesthetics while delivering enterprise-grade AI computing. These capabilities make the AIR-420 a future-ready, dependable core for high-traffic exhibitions, enhancing visitor experiences with immersive, real-time AI art interactions.
Benefits
- Enhanced Visitor Experience: Guests enjoy the AI-powered photo booth and take home unique artistic portraits, boosting interactivity and engagement.
- Museum Technology Upgrade: With the AIR-420 Edge AI solution, art evolves
-25/AIR-420_05-20251103174438.png)
