Media Mining & Open-source analysis

Thanks to its expertise in information and content processing, Bertin IT has been invited to collaborate on a number of different research projects dedicated to the investigation and enhancement of open-source, multilingual and multimedia (web,TV, radio) data

Media Mining & Open-source analysis 2017-03-16T10:52:12+00:00

IMM – Intégration Multimédia Multilingue (2013 – 2016)

Led by the SystemX Technological Research Institute, the IMM project is aimed at designing a test, development and evaluation platform for the analysis of multilingual multimedia content. Given the explosion in the amount of data that is produced and disseminated around the world today, the ultimate aim is to meet the need for high-performance tools that are capable of extracting exploitable knowledge from this deluge of information.

Under the theme of “Data & Decisions”, the primary applications of the IMM project are intended for the monitoring of crisis situations and conflicts (identification of troublemakers, surveillance of influence groups …) and for cyber-defence (detection of threats of attacks on critical information systems …).

Learn more about IMM

Bertin IT’s role in IMM

For the “Information Networks” element of the project, Bertin IT provides its MediaCentric® solution, along with its expertise in the in-depth analysis of multilingual multimedia open sources, with a view to designing and developing tools for collecting, indexing, analyzing and viewing data drawn from social networks, and all this in as close to real-time as possible.

A specialist in automatic voice processing, Vecsys provides key input on the major issue of accelerating the development of voice recognition engines (error analysis and collection of target data) and the adaptation and integration of these engines for searching for information contained in web-based multilingual texts and videos.

Vecsys also participates in the architecture design process and in the scalability tests of the SaaS infrastructure.

VIRTUOSO – Versatile InfoRmation Toolkit for end-Users oriented Open Sources exploitation (2009 – 2013)

With the backing of the European Commission (FP7), the aim of the VIRTUOSO project is to supply a set of open-source information processing tools to the European players in the field of Security.

These tools include applications ranging from data collection to decision-making aids, offering reliable information (evaluation and validation) at each stage of the process. The result of cooperation between the different European agencies, the tools are standard assessed in order to facilitate the exchange and storage of information.

Learn more about VIRTUOSO

Bertin IT’s role in VIRTUOSO

Bertin IT was assigned with the following tasks:

  • Multimedia Web acquisition system: Webcast-Tracker
  • Multimedia TV/Radio acquisition system: SAT-Tracker
  • Audio processing: Transcription, Speaker identification, Language Identification
  • Text processing: Machine Translation

All the above functions are to be found in MediaCentric®.

QUAERO (2008-2013)

A Franco-German collaborative research program launched in 2008 with the backing of Oséo Innovation, Quaero brought together more than 30 private & public partners for a series of application-based projects focusing on the automatic analysis, classification and use of multilingual multimedia content.

Banking on a high-level of interaction between the people from the worlds of research and industry, Quaero mobilized over 1000 engineers and research scientists of 35 different nationalities, specialists in the automatic processing of voice, text, image, video, audio and music content and also in data protection.

Learn more about QUAERO

QUAERO Demo Session Video and Music Search Engine (MSSE)

Bertin IT’s role in QUAERO

During the 5 year life of the project, Bertin IT and Vecsys contributed their know-how in detailed analysis of multimedia mutilingual content and automatic voice processing to a series of different projects. They made a particular contribution to the birth of MSSE, a prototype multimedia (audio/video) search engine developed in partnership with Orange, LTU, INA and IRCAM.

Bertin IT also perfected its MediaCentric® solution, enhancing the collection, correlation and analysis functions and adding a new results overview feature.

Vecsys was also closely involved in one of Quareo’s flagship projects, Technicolor’s Smart TV, via the design of the voice command system that provides users with immediate and intuitive access to their desired content from the mass of programmes, applications and services on offer.

Vecsys expertise is also found in Exalead’s CloudView demonstrator, which enables users to navigate their way through debates recorded at the National Assembly. Thanks to its word-text synchronization technology, a user can search for and listen back to a precise extract just by providing a quotation, a name or a key word for example.

HERISSON – Habile Extraction du Renseignement d’Intérêt Stratégique à partir de Sources Ouvertes Numérisées (2010 – 2012)

This upstream research program launched by the DGA, the French arms procurement agency, was aimed at evaluating commercial, off-the-shelf software (COTS) for the processing of open source data, i.e. data that is broadcast and freely available to all (as opposed to private data such as e-mails, P2P downloads and FTP file shares).

Some fifteen or so performance tests were conducted on behalf of operational units from the Intelligence Community along with a further twenty technical appraisals of over 50 software items that could be used in Open-Source Intelligence (OSINT).

Learn more about HERISSON

Bertin IT’s role in HERISSON

Besides the appraisal of the COTS and OS software, Bertin IT was also tasked with integrating the video, audio/voice, translation, OCR and TV/Radio acquisition software into the test platform.

During this work, Bertin IT was able to improve its understanding of the operational needs of the different entities of the Defence Ministry in terms of Open-Source Intelligence (OSINT). Appraisal of the diverse range of technological solutions on the market brought to light the pertinence of innovative processing techniques for OSINT. All of the positive findings from this research are now incorporated into the MediaCentric® platform.