Document AI
Discover our technology
Our processing capacities
Our AI model integrations
Our software Arkindex
Our processing capacities
TEKLIA provides a comprehensive range of artificial Intelligence technologies for document recognition, optimized to extract information from historical and cultural sources.
TEKLIA's expertise covers six different document processing tasks, that can be combined in a single workflow.
Automatic Text Recognition (OCR/HTR)
Extract printed or handwritten text from your documents.
Our Automatic Text Recognition capabilities allows for the conversion of scanned documents into fully searchable and editable text.
Document Layout Analysis
Segment, classify and link elements within a page.
Hierachise elements within a corpus.
Our layout analysis technology segments pages into logical regions (titles, paragraphs, images, and tables...),preserving visual hierarchy and context for more accurate content interpretation.
Table and Data Recognition
Capture and structure complex tabular data.
Our system detects and extracts tables, figures, and numerical data with layout-aware precision, enabling easy transformation into usable formats like CSV or Excel for further analysis.
Named-Entities Recognition
Identify and categorize key information automatically.
Our Named-Entities Recognition system detects names, dates, organizations, locations, and other critical entities within your documents, making content analysis faster and more structured.
Media Cataloging
Classify and index media automatically by content.
Our media cataloging solution automatically tags, classifies, and indexes image files, enabling augmented integration into digital archives.
Image Search by Content (CBIR)
Enhance the discoverability of your media.
Our similarity-based search allows you to discover visually related images by content rather than keywords, enhancing discovery and retrieval across large multimedia collections.
Our AI model integrations
We target each processing tasks by choosing the best performing model on the market and implementing it in our workflow.
We can integrate emerging AI algorithms and models into our Arkindex software, ensuring our customers have ongoing access to the latest advances and enabling them to continuously improve document processing performance.
There is a wide variety of AI models capable of performing the same task.
Thanks to our dual expertise in technology and the heritage sector, we can identify the most relevant models for processing your documents.
Our processing software
Arkindex
The open-source Arkindex
platform, developed by TEKLIA, allows you to structure and automate
heritage collection processing by combining automatic recognition
algorithms, manual processing, and the import of existing metadata.
- Designed to operate on millions of pages
- Integrates all types of documents (images, PDFs, IIIFs)
- Orchestrates complex chains of open-source or proprietary algorithms (OCR/HTR, object detection, classification, entity extraction, etc.)
- Ensures the traceability and replicability of processing..