Immersive reconstructions of archaeological sites often present visitors with richly modelled environments but little opportunity for personalised inquiry. In a collaboration between the University of Genoa and the 3D Lab Factory, we present a pilot that augments the VR installation of the Museo Preistorico dei Balzi Rossi (Ventimiglia, Italy) with an AI-enhanced guide modelled on the Upper-Palaeolithic "Dame du Cavillon." The avatar blends a large-language-model dialogue engine with MetaHuman embodiment in Unreal Engine 5.5, drawing on a curated archaeological knowledge bank to keep every utterance historically grounded. Spoken interaction, gestural cues, and guided movement are orchestrated in real time, allowing the guide to answer questions in Italian, English, and French, escort visitors to points of interest, and gracefully deflect anachronistic prompts. The result is both a narrative anchor and a conversational interface that personalises the learning journey and demonstrates how AI-driven VR can deepen public engagement with archaeological heritage.
Real-Time Multilingual Museum Guidance with ConvAI and Unreal Engine 5
Gaia Lorenzoni;Matteo Scaramuzzino;Saverio Iacono;Luca Martini;Daniele Zolezzi;Gianni Vercelli
2025-01-01
Abstract
Immersive reconstructions of archaeological sites often present visitors with richly modelled environments but little opportunity for personalised inquiry. In a collaboration between the University of Genoa and the 3D Lab Factory, we present a pilot that augments the VR installation of the Museo Preistorico dei Balzi Rossi (Ventimiglia, Italy) with an AI-enhanced guide modelled on the Upper-Palaeolithic "Dame du Cavillon." The avatar blends a large-language-model dialogue engine with MetaHuman embodiment in Unreal Engine 5.5, drawing on a curated archaeological knowledge bank to keep every utterance historically grounded. Spoken interaction, gestural cues, and guided movement are orchestrated in real time, allowing the guide to answer questions in Italian, English, and French, escort visitors to points of interest, and gracefully deflect anachronistic prompts. The result is both a narrative anchor and a conversational interface that personalises the learning journey and demonstrates how AI-driven VR can deepen public engagement with archaeological heritage.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.



