Book Finder is a web-based application that uses the Wimpole model, an AI object recognition system, to identify books from their spines without removing them from the shelves. It retrieves detailed metadata about each book and makes previously inaccessible historic collections available to the public through augmented reality.
This tool was developed as part of the Augmented Browsing of Books in Historic Libraries project led by the University of the Arts London (CCW Graduate School). The project aimed to address a common issue in historic houses and libraries, where books, despite being significant collections, are not accessible to visitors. A notable case is Wimpole Hall, which holds a fully catalogued but physically inaccessible book collection.
The solution involves mobile-based AR technology allowing users to hover their phone cameras in front of book spines to identify volumes and access detailed information without handling the physical objects.
More about the research project: UKRI Project Reference AH/R010102/1
- Identifies books via AI by analysing the spine without removing the book.
- Displays metadata and contextual information through an AR interface.
- Protects delicate historical books by enabling touchless access.
- Enhances visitor experiences in historic libraries and properties.
- Type: Webtool / Application
- AI Model: Wimpole model (object recognition)
- Year Produced: 2018
- Licence: Open Source
Book Finder allows members of the public to explore and engage with historic book collections that would otherwise remain unseen. It supports interpretation, education, and conservation by providing accessible digital interactions.
Key Partners:
- Ligatus, University of the Arts London
- Museotechniki
- National Trust
Key Stakeholders:
- Institute of English Studies, University of London
- University of Uppsala Library
- Bodleian Library, University of Oxford
- John Rylands Library, University of Manchester
- Salisbury Cathedral
- Christchurch College, Oxford