Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

Simon Fraser University

$46,000.00

Neural Field World Modeling

FY 2024-25

Summary
Humans build a rich representation of the world in their minds by observing and interacting with the world everyday. Computational systems equipped with such perception and representation capabilities would have tremendous impacts on our society. For example, home-robots would be able to build the map of a house and understand how the space is utilized for handling house-chores and taking care of the elderly. At construction sites, understanding building structures and their changes would enable automated progress monitoring and safety hazard detection, which could reduce 60 billion dollars worth of waste due to inefficiency and 4,000 lost lives of construction workers in North America every year. In the space of design, computer aided design (CAD) software could analyze day-to-day usage patterns of an artist and provide personalized help in speeding-up the design process, which would make the work of professional designers affordable to more consumers. Towards such computational intelligence, the last decade has witnessed the surge of deep neural networks (DNNs). State-of-the-art DNN models master the task of image understanding, for example, classifying the category of an image, detecting objects in an image, or segmenting out an object region from an image. The techniques achieve superhuman performance by absorbing a massive amount of pre-annotated training samples. This proposal will take this emerging field of computer vision to the next level by a family of three neural field representations: Departing from a mere task of single image understanding towards aggregating streams of perceptual signals to build and maintain consistent, compact, and efficient representations of the world. More specifically, the three research themes will be 1) Neural grid field of multi-modal sensor data for Indoor GPS; 2) Neural structured field of design variations for Autonomous CAD (Computer Aided Design); and 3) Neural dynamic structured field of action history for Scan2BIM (Scan to Building Information Model). The fourth research theme will optimize the latency, scalability, and energy-efficiency of these representations towards edge-computing applications. Besides advancing the state-of-the-art in 3D computer vision on many fronts, the research will be transformative to other academic disciplines such as robotics, machine learning, computer graphics, embodied AI, and civil engineering. The developed technologies will have immediate impacts on broader industries, ranging from visual-effects to gaming, digital-mapping, drones, autonomous cars, architecture, engineering and construction, where the Canadian visual effects industry is the third largest in the world, and Toronto, Montreal, and Vancouver boast fast-growing start-up ecosystems. Lastly, the research program will provide highly skilled engineers and scientists for Canada to remain competitive in the global marketplace.
Program
Discovery Grants Program - Individual
Location
British Columbia, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2023-04219
Name nom
Furukawa, Yasutaka
Department département
Computing Science
Organization id
5
Institution établissement
Simon Fraser University
Province en
British Columbia
Province fr
Colombie-Britannique
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2023
Award amount
46000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
801
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Computer software
Area of application fr
Logiciel
Research subject code
2800
Research subject group en
Artificial intelligence
Research subject group fr
Intelligence artificielle
Research subject en
Artificial intelligence (computer vision, use2603)
Research subject fr
Intelligence artificielle (vision artificielle, utiliser 2603)
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Neural Field World Modeling
Keywords
indoor mapping; indoor gps; 3d reconstruction; deep learning; augmented reality; virtual reality; localization; generative models; generation; 3d computer vision
Application summary
Humans build a rich representation of the world in their minds by observing and interacting with the world everyday. Computational systems equipped with such perception and representation capabilities would have tremendous impacts on our society. For example, home-robots would be able to build the map of a house and understand how the space is utilized for handling house-chores and taking care of the elderly. At construction sites, understanding building structures and their changes would enable automated progress monitoring and safety hazard detection, which could reduce 60 billion dollars worth of waste due to inefficiency and 4,000 lost lives of construction workers in North America every year. In the space of design, computer aided design (CAD) software could analyze day-to-day usage patterns of an artist and provide personalized help in speeding-up the design process, which would make the work of professional designers affordable to more consumers. Towards such computational intelligence, the last decade has witnessed the surge of deep neural networks (DNNs). State-of-the-art DNN models master the task of image understanding, for example, classifying the category of an image, detecting objects in an image, or segmenting out an object region from an image. The techniques achieve superhuman performance by absorbing a massive amount of pre-annotated training samples. This proposal will take this emerging field of computer vision to the next level by a family of three neural field representations: Departing from a mere task of single image understanding towards aggregating streams of perceptual signals to build and maintain consistent, compact, and efficient representations of the world. More specifically, the three research themes will be 1) Neural grid field of multi-modal sensor data for Indoor GPS; 2) Neural structured field of design variations for Autonomous CAD (Computer Aided Design); and 3) Neural dynamic structured field of action history for Scan2BIM (Scan to Building Information Model). The fourth research theme will optimize the latency, scalability, and energy-efficiency of these representations towards edge-computing applications. Besides advancing the state-of-the-art in 3D computer vision on many fronts, the research will be transformative to other academic disciplines such as robotics, machine learning, computer graphics, embodied AI, and civil engineering. The developed technologies will have immediate impacts on broader industries, ranging from visual-effects to gaming, digital-mapping, drones, autonomous cars, architecture, engineering and construction, where the Canadian visual effects industry is the third largest in the world, and Toronto, Montreal, and Vancouver boast fast-growing start-up ecosystems. Lastly, the research program will provide highly skilled engineers and scientists for Canada to remain competitive in the global marketplace.
Simon Fraser University - $46,000 Grant (2024-25) | Canada Spends