Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Engage Grants Program

École de technologie supérieure

$25,000.00

A Evaluation of Fusion Techniques for Face Recognition in Video Surveillance

EF 2014-15

Résumé
Automatic face recognition (FR) is increasingly employed by public safety organizations for enhanced securityand situational awareness. However, recognizing faces corresponding to target individuals (belonging to awatchlist) remains a challenging problem in video surveillance (VS) applications. Accuracy of state-of-the-artFR technologies declines due to variations of capture conditions (e.g., pose and illumination) in video streams.Moreover, FR must be rapid and scalable to the number of faces per frame, the frame rate, and the number ofcameras in a VS network. NextgenID Canada Inc. provides its customers with VS solutions for face screeningthat rely on well-positioned cameras and commercial FR technologies. For reliable FR under semi-controlledconditions, specialized techniques must presently be integrated with these technologies. Besides controllingconditions to capture higher quality faces, integrating techniques for fusion of the matching scores output by acommercial system over multiple templates, frames and cameras should greatly improve the accuracy androbustness of customer solutions.This project aims to provide NextgenID with the expertise needed to deploy FR systems with reliableinformation fusion. Advanced fusion techniques that are suitable for FR in watchlist screening applications willbe compared and recommended. Several techniques will be evaluated for fusion of matching scores producedby the Cognitec FaceVACS FR software. Faces captured over a network of video cameras will be matchedagainst several diverse reference images - high quality stills collected during enrollment with different poseangles, and levels of illumination and resolution. In particular, new Boolean Combination techniquesintroduced by the author allow for fusion of scores in the ROC space, and have been shown to outperformsome leading techniques, even when target individuals are rarely captured in a video. Experiments will beperformed using anonymized video data provided by NextgenID, and performance will be assessed in terms ofaccuracy and resource requirements.
Programme
Engage Grants Program
Emplacement
Quebec, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
463850-2014
Name nom
Granger, Eric
Department département
Génie de la production automatisée
Organization id
53
Institution établissement
École de technologie supérieure
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2014
Competition year année de concours
2014
Award amount
25000
Program id
EGP
Program name en
Engage Grants Program
Program name fr
Programme de subventions d'engagement partenariat
Group en
Research partnerships programs
Group fr
Programmes de partenariat
Committee code
1552
Committee name en
Quebec Internal Decision Committee
Committee name fr
Comité de décision interne pour le Québec
Area of application code
1207
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Information, computer and communication technologies
Area of application fr
Technologie de l'information, des ordinateurs et des réseaux de communications
Research subject code
2605
Research subject group en
Robotics
Research subject group fr
Robotique
Research subject en
Pattern analysis and machine intelligence
Research subject fr
Analyse de formes et intelligence artificielle
Partie
1
Nb partie
1
Application title
A Evaluation of Fusion Techniques for Face Recognition in Video Surveillance
Keyword
NSERC
Application summary
Automatic face recognition (FR) is increasingly employed by public safety organizations for enhanced securityand situational awareness. However, recognizing faces corresponding to target individuals (belonging to awatchlist) remains a challenging problem in video surveillance (VS) applications. Accuracy of state-of-the-artFR technologies declines due to variations of capture conditions (e.g., pose and illumination) in video streams.Moreover, FR must be rapid and scalable to the number of faces per frame, the frame rate, and the number ofcameras in a VS network. NextgenID Canada Inc. provides its customers with VS solutions for face screeningthat rely on well-positioned cameras and commercial FR technologies. For reliable FR under semi-controlledconditions, specialized techniques must presently be integrated with these technologies. Besides controllingconditions to capture higher quality faces, integrating techniques for fusion of the matching scores output by acommercial system over multiple templates, frames and cameras should greatly improve the accuracy androbustness of customer solutions.This project aims to provide NextgenID with the expertise needed to deploy FR systems with reliableinformation fusion. Advanced fusion techniques that are suitable for FR in watchlist screening applications willbe compared and recommended. Several techniques will be evaluated for fusion of matching scores producedby the Cognitec FaceVACS FR software. Faces captured over a network of video cameras will be matchedagainst several diverse reference images - high quality stills collected during enrollment with different poseangles, and levels of illumination and resolution. In particular, new Boolean Combination techniquesintroduced by the author allow for fusion of scores in the ROC space, and have been shown to outperformsome leading techniques, even when target individuals are rarely captured in a video. Experiments will beperformed using anonymized video data provided by NextgenID, and performance will be assessed in terms ofaccuracy and resource requirements.