François Brémond

Contents:

Spatio-Temporal Activity Recognition Systems

Research topics:

Computer Vision: video processing, object detection and tracking, motion analysis, pattern recognition,

Cognitive Vision: video understanding, scene understanding, event recognition, behaviour analysis, activity monitoring, muti-sensor fusion, multimedia interpretation,

Machine Learning: deep learning architectures, data augmentation, self-attention, knowledge distillation, contrastive learning, self-learning, life long learning, knowledge-based systems, spatio-temporal reasoning, uncertainty handling,

Autonomous Systems: real-time systems, system evaluation, parameter tuning, system design, 3D visualisation,

Resume:

François Brémond is a Research Director DR1 at INRIA Sophia Antipolis. He created the STARS team on the 1st of January 2012 and was previously the head of the PULSAR INRIA team in September 2009. He obtained his Master degree in 1992 from ENS Lyon. He has conducted research work in video understanding since 1993 both at Sophia-Antipolis and at USC (University of Southern California), LA. In 1997 he obtained his PhD degree from INRIA in video understanding and François Brémond pursued his research work as a post doctorate at USC on the interpretation of videos taken from UAV (Unmanned Airborne Vehicle) in DARPA project VSAM (Visual Surveillance and Activity Monitoring). In 2007 he obtained his HDR degree (Habilitation à Diriger des Recherches) from Nice University on Scene Understanding: perception, multi-sensor fusion, spatio-temporal reasoning and activity recognition. He is a co-fonder of Keeneo, Ekinnox and Neosensys, three companies in intelligent video monitoring and business intelligence. He also co-founded the CoBTek team from Nice University on the 1st of January 2012 with P. Robert from Nice Hospital on the study of behavioral disorders for older adults suffering from dementia.

He designs and develops generic systems for dynamic scene interpretation. The targeted class of applications is the automatic interpretation of indoor and outdoor scenes observed by sensors and in particular by monocular colour cameras. These systems detect and track mobile objects, which can be either humans or vehicles, and recognize their behaviours. He is particularly interested in filling the gap between sensor information (pixel level) and behaviour recognition (semantic level). François Brémond is author or co-author of more than 200 scientific papers published in international journals or conferences in video understanding. He is reviewer for several international journals (CVIU, IJPRAI, IJHCS, PAMI, AIJ, Eurasip JASP,...) and conferences (CVPR, ICCV, AVSS, VS, ICVS,…). He has (co-)supervised 18 PhD theses. He is an EC INFSO and French ANR Expert for reviewing project. He was teaching numerical classification at Nice University and video understanding in a High Engineer School at Master level.

He has participated to 12 European projects (Esprit, ITEA, FP6, FP7: PASSWORDS, ADVISOR, AVITRACK, SERKET, CARETAKER, CANTATA, COFRIEND, SERKET, VICOMO, VANAHEIM, SUPPORT, DEM@CARE), one DARPA project, 12 French projects (ANR, DGE, Prédit, TechnoVision, PACA, CG06,...), several industrial research contracts (Bull, Vigitec, SNCF, RATP, ALSTOM, STMicroElectronics, Thales, Keeneo, LinkCareServices, Neosensys...) and several international cooperations (USA, Taiwan, UK, Belgium) in video understanding. For instance, he has succeeded to recognize a large variety of scenarios in different applications: fighting, abandoned luggage, graffiti, fraud, crowd behavior in metro stations, in streets and onboard trains, aircraft arrival, aircraft refuelling, luggage loading/unloading on airport aprons, bank attack in bank agencies, access control in buildings, office behavior monitoring for ambient intelligence, older adult activity monitoring for homecare applications and wasp monitoring for biological application. He has also participated to a series of ARDA workshops to build an ontology of video events.

See other documents:

En Français:

François Brémond est directeur de recherches DR à INRIA Sophia Antipolis (DR2 depuis le 1er octobre 2010, DR1 depuis le 1er janvier 2018). Il a créé l'équipe projet INRIA STARS et il est membre fondateur de l'équipe CoBTeK de l'université de Nice avec le professeur P. Robert du CHU de Nice sur l'étude des troubles du comportement pour les personnes âgées. François Brémond a mené ses travaux de recherche dans le domaine de l'interprétation dynamique de scène à partir de données capteurs depuis 1993, à Sophia-Antipolis et à l'USC (University of Southern California). Il s'intéresse plus particulièrement à construire un modèle numérique des comportements humains. Il a participé à 12 projets européens sur l'analyse et l'interprétation vidéo (Esprit, ITEA, FP6, FP7), à 12 projets nationaux (ANR, DGE, Prédit, PACA, CG06,...), à de nombreux contrats industriels (Bull, Vigitec, SNCF, RATP, ALSTOM, STMicroElectronics, Thales,...) et plusieurs coopérations internationales (USA, Taiwan, UK, Belgique, Singapour). Il a conçu des systèmes de reconnaissance de scénarios au cours de différentes applications de vidéo surveillance et monitoring : bagarre, bagage abandonné, graffiti, fraude, comportements anormaux de groupes d'individus et foule dans des stations de métro, dans des espaces publics et dans des trains, arrivage et maintenance d'avion (e.g. plein d'essence), chargement et déchargement de bagages dans un avion, surveillance d'agence bancaire, contrôle d'accès dans des bâtiments, ainsi que le monitoring de personnes âgées pour des applications de soins à domicile et le monitoring de guêpes pour une application biologique. Il est (co-)auteur de plus de 200 papiers scientifiques et a (co-)encadré plus de 18 thèses de doctorat. Il est également co-fondateur de la start-up Keeneo/Digital Barriers, spécialiste de la vidéo protection intelligente et de la start-up Neosensys, spécialiste de la Business Intelligence dans la grande distribution. Il a servi en tant qu'expert auprès de l'ANR et de la CEE pour le suivi de projets européens FP7.
Domaines de recherche:

Research Projects:

For more information, see the complete list of 43 Research Projects