menu
R&D Project Manager
Digital Intelligence
As Project Manager, I'm in charge of advanced R&D programs and projects around multimodal & multilingual analysis.
Main skills implemented in projects : Data Acquisition (SAT, OTA, Web), Multimodal & Multilingual Analysis, Video Summarization, Face Detection & Recognition, Deep Metadata Extraction, Image & Video Retrieval, OCR, Spoken Language Identification, Speaker Identification, Automatic Speech Recognition, Machine Translation, Information Retrieval, NLP, Knowledge Management.
Advanced research programs for French MoD / Direction Générale de l'Armement (PEAs)
- Large scale platform dedicated to the processing of audio flows
- Experimental OSINT platform dedicated to evaluation and benchmarking of COTS & OSS
- Multimedia Services for an OSINT platform (video, speech, OCR, MT)
QUAERO European R&D program / OSEO Innovation
- Online Multimedia Translation Platform (OMTP - Leader : SYSTRAN)
- Multimedia Search Services to enrich European portals (MSSE - Leader : Orange Labs)
- Customer Media Search Engine (CMSE - Leader : Exalead/Dassault Systèmes)
European Union FP7 Security project
Advanced research programs (PEAs) for French MoD / Direction Générale de l'Armement
Large scale platform dedicated to the processing of audio flows of all kinds, from open sources (radio, television, podcast, etc...) to informal conversation, for monitoring and alert purposes ;
 |
 |
|
 |
 |
|
 |
||
Experimental OSINT platform dedicated to evaluation and benchmarking, according to technical and usability factors, of Open-Source Software & Commercial Off-The-Shelf software.
 |
|
|
|
|
Provider of processing services for video, audio, image and text for an OSINT platform : Data Acquisition (SAT, OTA, Web), Video Summarization, Face Detection & Recognition, OCR, Spoken Language Identification, Speaker Identification, Automatic Speech Recognition and Machine Translation.
 |
 |
 |
 |
 |
 |
 |
|
 |
 |
 |
 |
QUAERO R&D program / OSEO Innovation
Innovative developments with laboratories (RWTH, LIMSI/CNRS, KIT, ...) in particular for the translation of vocal data. Provider of processing services for video, audio and text : Multimedia Data Acquisition (Web Mining), Face Detection & Recognition, Automatic Speech Recognition, Machine Translation.
 |
 |
|
| |
|
|
 |
 |
Bertin IT had developed an adaptive platform, called MediaCentric®, to acquire, filter, organize and manipulate large collections of open source multimedia and multilingual documents. Bertin IT provides its advanced solution for multimedia multilingual content processing as platform for the demonstration of the Quaero CTC Technologies.
Inside Quaero program, I'm in charge of 3 application projects OMTP, MSSE and CMSE described below.
Online Multimedia Translation Platform (OMTP - Leader : SYSTRAN)
The goal of this project is to build a platform of next generation translation tools. Serving as a pure showroom, this platform will help to demonstrate how an internet user can easily translate any kind of multimedia contents such as video, audio, RSS flow, instant messaging...
I developed the OMTP portal and platform including Multilingual Multimedia Search taking advantage of Speech to Text (STT) technologies applied to audio streams (from pod casts, videos ...) coupled to Machine Translation (MT) technologies would bring more results to the user, and therefore increase the search engine overall relevancy.
The application dedicated to “Multilingual Multimedia Search” is build on top of the MediaCentric® platform.
Multimedia Search Services to enrich European portals (MSSE - Leader : Orange Labs)
This project aims to develop a searching and navigation service prototype allowing to access to various audiovisual contents.
The major innovations lie in the use of advanced content analysis technologies which enrich automatically their description, and thus improve the relevance of the results and the user experience. MSSE addresses video contents in several languages: French, English, German, Spanish, Portuguese, Russian, Chinese and Arabic.
To allow a French user (for instance) to access contents in other languages, I integrated a Cross-Language Information Retrieval function based on machine translation of metadata (video titles and summaries) and on translating texts resulting from speech to text transcriptions. Indexation of metadata and speech translated is operated in French language in order to meet French queries. To maximize the efficiency of machine translation, we use two machine translation systems: the Bertin statistical MT system and the Systran rule-based (hybrid) one. The Bertin MT system is based on the RWTH (Rheinisch-Westfälische Technische Hochschule Aachen) MT technology.
The application dedicated to “Multimedia Cross-language information retrieval (M-CLIR)” is build on top of the MediaCentric® platform.
Customer Media Search Engine (CMSE - Leader : Exalead/Dassault Systèmes)
“Learning Analytics”
Build an "Educational data mining" application is one of aims of the CMSE project. I developed a sub-system for acquisition of educational video from Internet sources. The focus was given to discovery of podcasts available on web from educational video websites. We have targeted academic websites such as MIT, University of California Television and so on ... or multidisciplinary conference websites such as TED (Technology, Entertainment and Design).
To provide ability to analyse educational video contents, I developed automatic module for video text detection and recognition in multimodal context. In order to perform a deep educational video analysis, the module determines where text is visible (slides, transparency, …) in the video using an unsupervised mode relying on results correlation of processing.
Finally, I realised of complete chain including multimedia data acquisition, video and image processing and intuitive interfaces for visualizing and interacting with educational video. Beyond “Educational data mining”, video text detection and recognition can be used in many applications, such as semantic video indexing, summarization, multilingual video information access, etc.
The application dedicated to “Educational data mining” is build on top of the MediaCentric® platform.
European Union FP7 Security project
VIRTUOSO Project
Provider of processing services for video, audio and text for an OSINT platform : Multimedia Data Acquisition Systems (Web & SAT), Automatic Speech Recognition, Machine Translation.
 |
|
|
|
||
VIRTUOSO is a European project co-funded by the European Union under its FP7 Programme. The objective of the project is to develop a technical platform for the integration of tools for collection, processing, analysis, visualization and decision-support of open source information.
This middleware platform enables "plug-and-play" functionalities that will improve the ability of open source information professionals to utilize data from across the source and format spectrum in support of the decision-making process. Upon completion of the project, the VIRTUOSO platform will be released as open-source-code software.
As a proof of concept and to highlight the efficiency of the VIRTUOSO platform, a demonstrator with functionalities provided by the consortium partners is being built. This demonstrator is validated using operational scenarios.
The VIRTUOSO project complies with legal regulations and enforces the principles of privacy and data protection. By doing so, the project ensures that the interests of the citizens of the European Union are being respected.
The parts of VIRTUOSO project dedicated to “Mulimedia data acquisition and speech-to-text processing” are build on top of the MediaCentric® platform.