Description
MetaAssistant is an AI-based automatic cataloging platform designed to support catalogers and archivists in describing large quantities of images.
The system analyzes each photograph, recognizes people, places, objects, and context, and generates structured metadata and a description consistent with archival and cataloging standards.
Features
- Automatic import from digital libraries via API
- Face detection in each image
- Subject recognition through celebrity databases and internal archives
- Incremental learning to improve recognition of unknown faces
- Automatic generation of content and context descriptions
Architecture and Technologies
MetaAssistant is a modular cloud platform capable of managing the automatic ingestion of images and orchestrating multiple artificial intelligence engines specialized in:
- Facial recognition: neural networks for face detection and face recognition on mixed databases
- Semantic analysis: interpretation of relationships, poses, actions, costumes, scenography, and historical context
- Metadata generation: composition of collected data into structured records and production of archival descriptions
Advantages
- Significant reduction in cataloging time compared to manual work
- Scalability for very large archives
- Higher metadata quality, including roles, relationships between subjects, historical context, and framing
- Improved data usability, thanks to more relevant and consistent descriptions
Use Cases
- Mass cataloging of historical photographic archives
- Support for archival digitization projects
- Generation of consistent archival descriptions for digital portals
- Advanced analysis to recover historical context, relationships between subjects, and temporal information