Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Engage Grants Program

University of Waterloo

$25,000.00

Software driven early fault detection in large scale machine learning systems

FY 2016-17

Summary
Miovision Technologies is a Canadian company providing intelligent traffic solutions to hundreds of customers from over 37 countries. Miovision is currently rolling out a new product line, called Connected Intersection, designed to connect traffic cabinets to transportation control centres via mobile networks. Connected Intersection is capable of monitoring traffic and signals as well as monitoring live video from the intersections. The company is incorporating automatic scene understanding using computer vision and machine learning techniques to classify live traffic events. Their video analytic software employs a supervised machine learning software and a large scale traffic dataset to learn traffic scenes, vehicles, pedestrians, and bikes. Then, the video analytic software is applied for traffic surveillance applications and to convert customer's traffic video data into traffic reports. Given the variety of camera measurement noise caused due to snow, fog, rain, change in lighting, and static and dynamic occlusions, developing a video analytic software that provides human-level accuracy is extremely difficult. Under abnormal weather conditions, the performance of the current video detection system is highly unpredictable. Therefore, currently, they employ an expensive human-based mechanism to determine failure modes. The goal of this project is to design a software driven fault identification system by employing the concept of mutation testing. In the short-term, we will achieve two key milestones: (M1) develop physical models for rain, snow, and night sky to synthesize a variety of noisy images in rainy, snowy and cloudy conditions; and (M2) employ the mutation testing technique to evaluate the performance of machine learning based video analytic algorithms by using the models developed in M1. The proposed project and collaboration will help increase Miovision's competitiveness on an international scale, therefore leading to substantial growth of this industry in Canada. Our long-term objective beyond this project is to develop new methodologies for software testing by applying machine learning techniques.
Program
Engage Grants Program
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
493707-2015
Name nom
Naik, Kshirasagar
Department département
Electrical & Computer Engineering
Organization id
33
Institution établissement
University of Waterloo
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2016
Competition year année de concours
2015
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
1553
Committee name en
Ontario Internal Decision Committee
Committee name fr
Comité de décision interne pour l'Ontario
Area of application code
700
Area of application group en
Transportation systems and services
Area of application group fr
Services et systèmes de transport
Area of application en
Transportation systems and services
Area of application fr
Services et systèmes de transport
Research subject code
2700
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Information technology
Research subject fr
Technologie de l'information
Partie
1
Nb partie
1
Application title
Software driven early fault detection in large scale machine learning systems
Keyword
NSERC
Application summary
Miovision Technologies is a Canadian company providing intelligent traffic solutions to hundreds of customers from over 37 countries. Miovision is currently rolling out a new product line, called Connected Intersection, designed to connect traffic cabinets to transportation control centres via mobile networks. Connected Intersection is capable of monitoring traffic and signals as well as monitoring live video from the intersections. The company is incorporating automatic scene understanding using computer vision and machine learning techniques to classify live traffic events. Their video analytic software employs a supervised machine learning software and a large scale traffic dataset to learn traffic scenes, vehicles, pedestrians, and bikes. Then, the video analytic software is applied for traffic surveillance applications and to convert customer's traffic video data into traffic reports. Given the variety of camera measurement noise caused due to snow, fog, rain, change in lighting, and static and dynamic occlusions, developing a video analytic software that provides human-level accuracy is extremely difficult. Under abnormal weather conditions, the performance of the current video detection system is highly unpredictable. Therefore, currently, they employ an expensive human-based mechanism to determine failure modes. The goal of this project is to design a software driven fault identification system by employing the concept of mutation testing. In the short-term, we will achieve two key milestones: (M1) develop physical models for rain, snow, and night sky to synthesize a variety of noisy images in rainy, snowy and cloudy conditions; and (M2) employ the mutation testing technique to evaluate the performance of machine learning based video analytic algorithms by using the models developed in M1. The proposed project and collaboration will help increase Miovision's competitiveness on an international scale, therefore leading to substantial growth of this industry in Canada. Our long-term objective beyond this project is to develop new methodologies for software testing by applying machine learning techniques.