Exploring the Potential of Federated Learning CNN for Interactive Virtual Tours of UNESCO Cultural Heritage Sites: A Case Study

A federated learning CNN model demonstrates high accuracy (95.4%) for item recognition and categorization in UNESCO World Heritage Sites, offering enhanced personalization and privacy compared to centralized models. The study trained the model using 40,000 images from four sites—Angkor Wat, Stonehenge, Machu Picchu, and Petra—employing the Federated Averaging technique with MobileNet CNN architecture in a simulated federated learning environment with 20 devices. Participants appreciated the model's ability to customize tours based on individual interests while maintaining robust privacy and security features, suggesting potential for broader adoption in virtual heritage exploration.

Mehta, S., Kukreja, V., & Gupta, R. (2023). Exploring the Potential of Federated Learning CNN for Interactive Virtual Tours of UNESCO Cultural Heritage Sites: A Case Study. 2023 14th International Conference on Computing Communication and Networking Technologies (ICCCNT), 1–6. https://doi.org/10.1109/icccnt56998.2023.10307071