Grant: Engage Grants Program
Concordia University
$25,000.00
Fractional Brownian motion based voice activity detection with application to subtitle failure detection for broadcasting services
FY 2012-13
Summary
Voice activity detection (VAD) has achieved significant research progress over the past two decades. However,
two issues become more notable when VAD is real-time implemented in commercial products: (1) Highly
unpredictable noise interferences often occur in realistic application scenarios. (2) Computational complexity
becomes a bottleneck preventing VAD from being used in low-cost commercial products. Although many
VAD methods have been proposed in speech processing research community, they have not yet been proved
successful in TV broadcasting industry. In this project, we will design a VAD solution, that is robust to
nonstationary noise interferences and has a low computational complexity, to the subtitle failure detection
problem in TV broadcasting products of Miranda Technologies.
Miranda Technologies, headquartered in Montreal, is a world-leading company that provides technical
solutions for TV broadcasting products and services. Miranda has recently found that one of the serious issues
experienced by its tech users is the subtitle failure. Subtitles are delivered to TV program viewers along with
audio and video stream data by a television content provider. However, the delivered subtitle data often
experiences a certain content failure, including occasional subtitle missing, out of synchronization between
subtitles and audio contents, and subtitle content mismatch like language mismatch and text mismatch. In this
project, we will design and implement a VAD-based subtitle failure detection (SFD) technique for Miranda's
TV broadcasting products. The proposed work has a direct relevance to the monitoring and control product line
of the supporting company. One of the Miranda's strategic goals is to broaden its product development
capabilities with automatic SFD, which can be made possible through the deployment of the VAD algorithm to
be developed in this project. The new technology is expected to contribute to Miranda by strengthening its
leadership role and market share in TV broadcasting industries worldwide.
Program
Engage Grants Program
Recipient
Location
Quebec, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
437992-2012
Name nom
Zhu, WeiPing
Department département
Electrical and Computer Engineering
Organization id
70
Institution établissement
Concordia University
Province en
Québec
Province fr
Québec
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2012
Competition year année de concours
2012
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
800
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information and communication services
Area of application fr
Services d'information et de télécommunications
Research subject code
2514
Research subject group en
Electrical and electronic engineering
Research subject group fr
Génie électrique et électronique
Research subject en
Digital signal processing
Research subject fr
Traitement des signaux numériques
Partie
1
Nb partie
1
Application title
Fractional Brownian motion based voice activity detection with application to subtitle failure detection for broadcasting services
Keyword
NSERC
Application summary
Voice activity detection (VAD) has achieved significant research progress over the past two decades. However,
two issues become more notable when VAD is real-time implemented in commercial products: (1) Highly
unpredictable noise interferences often occur in realistic application scenarios. (2) Computational complexity
becomes a bottleneck preventing VAD from being used in low-cost commercial products. Although many
VAD methods have been proposed in speech processing research community, they have not yet been proved
successful in TV broadcasting industry. In this project, we will design a VAD solution, that is robust to
nonstationary noise interferences and has a low computational complexity, to the subtitle failure detection
problem in TV broadcasting products of Miranda Technologies.
Miranda Technologies, headquartered in Montreal, is a world-leading company that provides technical
solutions for TV broadcasting products and services. Miranda has recently found that one of the serious issues
experienced by its tech users is the subtitle failure. Subtitles are delivered to TV program viewers along with
audio and video stream data by a television content provider. However, the delivered subtitle data often
experiences a certain content failure, including occasional subtitle missing, out of synchronization between
subtitles and audio contents, and subtitle content mismatch like language mismatch and text mismatch. In this
project, we will design and implement a VAD-based subtitle failure detection (SFD) technique for Miranda's
TV broadcasting products. The proposed work has a direct relevance to the monitoring and control product line
of the supporting company. One of the Miranda's strategic goals is to broaden its product development
capabilities with automatic SFD, which can be made possible through the deployment of the VAD algorithm to
be developed in this project. The new technology is expected to contribute to Miranda by strengthening its
leadership role and market share in TV broadcasting industries worldwide.