2024

Improving Reproducibility in Handwriting Text Recognition Research

This blog post explores the significance of reproducibility in Automatic Text Recognition (ATR) research, emphasizing its role in validating findings and promoting transparency within the ATR community. It also highlights key steps, such as open data access and standardized metrics, to ensure reproducibility and drive progress in the dynamic field of ATR.

By Christopher Kermorvant

Research
January 2024
2023

Discover TEKLIA's AI-Powered Photographic Analysis Service

Discover TEKLIA's game-changing service - a state-of-the-art AI-powered tool designed to transform the way museums, archives, and art institutions manage and analyze their photographic and illustrative collections.

By Christopher Kermorvant

Tech
November 2023

Colonial conscription and its role in the evolution of contemporary Mali: an exploration using automatic document recognition

Discover how the application of document recognition technology to colonial conscription registers provides access to previously unpublished information and opens up new perspectives for studying Mali's past.

By Christopher Kermorvant

Project
October 2023

TEKLIA to Present Four Research Papers at ICDAR 2023

TEKLIA is set to present four research papers at the ICDAR 2023, held in San José, California.

By Christopher Kermorvant

Research
August 2023

Beyond Handwritten Text Recognition

This article explores our unique approach of using deep learning to not only transcribe historical documents, but also extract valuable information and recognize complex handwritten tables. Learn how TEKLIA is revolutionising the processing of historical and patrimonial documents, dramatically simplifying research and improving the accessibility of historical data.

By Christopher Kermorvant

July 2023

Open-sourcing Callico

Callico, our crowdsourced annotation platform, is now released as open source software under the AGPL-v3 license!

By Bastien Abadie

Callico
July 2023

ICRC - Collaborative transcription and handwriting recognition applied to prisoners' handwritten lists

The International Committee of the Red Cross and Teklia are working on a new workflow for recognizing handwritten lists.

May 2023

AI for cataloguing at the Sainte Geneviève library

The Sainte Geneviève Library and Teklia have collaborated to perform automatic handwriting recognition on a digitised paper catalogue containing more than 50,000 records and 6,000 pages.

May 2023

Callico: A Comprehensive Solution for Document Annotation

Are you in search of a versatile and efficient document annotation tool? Look no further than Callico, a dedicated platform that outperforms its competitors in various aspects.

By Christopher Kermorvant

April 2023

Spacy is now fully integrated in Arkindex

Spacy, the Industrial-Strength Natural Language Processing tool, is now fully integrated into Arkindex for named-entity recognition. All the models available from spacy can be configured and applied on your documents in Arkindex.

By Christopher Kermorvant

March 2023
2022

New models, new possibilities for extracting information from digitised documents

By Christopher Kermorvant

December 2022

Automatic Text Recognition - The convergence between OCR and HTR technologies

By Christopher Kermorvant

December 2022

Belfort city archives: a pilot project for automatic recognition of city council registers

In collaboration with Teklia, the Archives of the City of Belfort have launched a pilot project consisting of the automatic transcription of all the registers of deliberations of the city councils.

By Christopher Kermorvant

November 2022

OCAPI: Outil de Captcha et d'Annotation du Patrimoine en Image

Découvrez comment sécuriser l'accès aux sites web tout en participant à l'indexation du patrimoine culturel.

By Christopher Kermorvant

October 2022

Meet Callico, Teklia's new collaborative platform for your document annotation campaigns

TEKLIA has designed and developed a collaborative web annotation platform with five annotation modes dedicated to documents recognition projects.

By Bastien Abadie

September 2022

When Automatic Document Processing meets Egyptian History

To celebrate the centenary of an important archeological campaign, Teklia has been selected by the IFAO to provide a platform allowing both the training of Deep Learning models for HTR on scanned pages, and the classification and indexation of data.

By Christopher Kermorvant

September 2022

SIMARA: automatic conversion of finding aids with handwriting recognition

TEKLIA was chosen by the French National Archives to develop a web application dedicated to the conversion of scanned handwritten finding aids, based on high-performance Deep Learning models for handwriting recognition.

By Christopher Kermorvant

July 2022

TEKLIA supports the 2022 Document Analysis Systems conference and will present two research papers

TEKLIA is an official sponsor of the 2022 DAS Conference and will present its latest research results in HTR and NER.

By Christopher Kermorvant

March 2022

BALSAC project registers have been processed!

An overview of our work on processing about 2 million scanned images in the BALSAC project

By Martin Maarand

February 2022

The Doc-UFCN library is now accessible to everyone!

A Python 3 library that allows you to apply Doc-UFCN models on your documents.

By Mélodie Boillet

January 2022

PDF and ALTO exports from Arkindex

A new version of our CLI tool lets you export an Arkindex project to PDF and ALTO files.

By Erwan Rouchet

January 2022

How IIIF 3.0 might affect non-compliance issues

An overview of some of the changes that IIIF 3.0 brings, and what could force non-compliant servers to change.

By Erwan Rouchet

January 2022
2021

Handling non-compliant IIIF servers in Arkindex

An overview of the many ways we work around common issues with external IIIF servers in Arkindex.

By Erwan Rouchet

December 2021

Validating your IIIF compliance automatically

An introduction to simple tools to ensure your server complies with IIIF.

By Erwan Rouchet

December 2021

What not to do when implementing IIIF

We ran into numerous issues while interacting with IIIF servers. This post documents some of them so that they can be avoided in the future or managed by other clients.

By Erwan Rouchet

November 2021

Ocelus now available for Arabic documents

TEKLIA releases a new version of Ocelus for automatic recognition of handwritten documents in Arabic.

By Marie Amyot

November 2021

Improving the performance of IIIF servers

A summary of Teklia's work on IIIF servers deployment choices in a performance objective.

By Valentin Rigal

November 2021

Open-sourcing our Transkribus client and PAGE XML parser

TEKLIA releases a tool to make interacting with Transkribus and PAGE XML in Python easier.

By Erwan Rouchet

October 2021

TEKLIA joins the SYNTHESYS+ European project

TEKLIA joins the Synthesys+ project to develop a platform for automatic data extraction from natural history specimen images

By Christopher Kermorvant

October 2021

Automatic recognition of 100 years of French Census: the SOCFACE project

The French National Research Agency funds a project by TEKLIA and its partners to process 100 years of French census.

By Marie Amyot

September 2021

Benchmark of IIIF servers for Machine Learning workflows

Performance analysis of several widely used open source servers implementing the IIIF API 2.0 specification, in the context of high-throughput Machine Learning workflows

By Valentin Rigal

June 2021

TEKLIA and Geosophy collaborate to extract information from geological archives

Geosophy and TEKLIA are very pleased to collaborate to demonstrate that archival documents can contain useful data to address contemporary issues such as environment, energy and housing.

By Christopher Kermorvant

April 2021

MACADAMIA: clinical information extraction to improve the care of newborns with congenital malformations

TEKLIA will develop Machine Learning models to assist medical doctors in extracting information from clinical report in order to create a large scale research database.

By Christopher Kermorvant

April 2021

What is the best export format for handwritten document processing results?

In the context of TEKLIA's involvement in the SYNTHESYS+ project, we took a look at 4 potential data formats to output the results of layout analysis, OCR and HTR, and NER processing.

By Marie-Laurence Bonhomme

April 2021

Arkindex presentation at the IIIF360 consortium workshop

Arkindex was presented at the IIIF event organized by the IIIF360 consortium (Biblissima, Campus Condorcet, TGIR Huma-Num)

By Christopher Kermorvant

March 2021
2020

TEKLIA collaborates with the Egyptian Ministry of Communications and Information Technology to implement AI-Powered projects

TEKLIA is very happy to announce a collaboration with the Ministry of Communications and Information Technology (MCIT) in Egypt to develop AI-based technology to automatically understand handwritten historical documents in Arabic.

By Christopher Kermorvant

September 2020
2019

TEKLIA will assist the Paris School of Economics in studying administrative documents in Nepali

TEKLIA will assist PSE in a research project dedicated to the study of local governance of forest resources in Nepal thanks to automatic information extraction from administrative records in Nepali

By Christopher Kermorvant

October 2019

TEKLIA presents two research papers at ICDAR2019

TEKLIA will present two research papers at the 15th International Conference on Document Analysis and Recognition organised by University of Technology Sydney (UTS), Australia and hold at the International Convention Centre (ICC) Sydney.

By Christopher Kermorvant

September 2019

TEKLIA's CaptchAN project awarded a grant from the Ministry of Culture

The CaptchAN project will develop a captcha service using cultural and patrimonial data provided by partner cultural institutions.

By Christopher Kermorvant

June 2019

TEKLIA joins the BALSAC project

TEKLIA will handle the automated transcription, named entity recognition and extraction from over 6 million digitized parish register entries (birth/baptism and death records mainly), dating from 1850 to 1920.

By Christopher Kermorvant

April 2019
2018

Emanuela Boros Thesis Defense

By Christopher Kermorvant

September 2018