Projects
Collaborating with Dapasoft Inc (external link, opens in new window) . and the Mount Sinai Hospital (external link) , we will develop interpretable machine learning algorithms for computer-aided diagnosis of Neonatal lung and heart diseases. Mount Sinai's Neonatal Intensive Care Unit is the largest in Canada, and point-of-care-ultrasound (POCUS) is the most common imaging modality for heart and lung disease diagnosis. However, highly specialized skillset is required for operating POCUS on neonates. We will also develop AI-guided training tools for new ultrasound technologists in this project.
Collaborating with Shaftesbury VR (external link) , we will develop a machine learning model for mutlimodal assessment of stress. The multimodal sensors can be physiological (e.g. EEG, heart-rate) and behavioural (e.g. facial expressions). The target is to use the assessed stress for Shaftesbury's Positive Distraction Entertainment System which adapts game content dynamically to reduce stress in children before a complex medical procedure, which can reduce complexity and recovery time.
Collaborating with Dapasoft Inc. (external link, opens in new window) , this project aims to develop new information processing tools and techniques to enhance emergency medical service. The target is to develop healthcare systems utilizing various IoT sensors to record the symptoms of the patients and analyze the data with machine learning and immersive context-aware visualization methods to assist the healthcare professionals in performing timely decision making in emergency situations.
Collaborating with AWE Company (external link) , this project aims at developing machine learning algorithms that can leverage point cloud data obtained from modern mobile AR libraries (ARKit/ARCore) to provide an intelligent understanding of the scene surrounding a user through 3D object recognition. The intelligent understanding will further enable collaborative augmented reality experiences between multiple users.
Collaborating with Alcohol Countermeasure Systems (external link) , we will investigate a novel cloud-based multimodal biometrics framework using multimodal data to address the issues of continuous authentication and drowsiness detection. The deliverables aim at developing novel techniques for effective driving safety devices in order to provide a reliable solution for the ever-increasing needs of safety and security on Canada's highway system.
Collaborating with SimentIT, we will develop new cloud-based mobile AR technologies for a museum exhibit. The project is in partnership with the Canadian Science and Technology Museum (CSTM) (external link) in Ottawa. The developed technology will be deployed in CSTM’s “artifact alley”, where users can experience a barrier-free and natural AR experience. Significant research challenge is real-time 3D object recognition and pose estimation under challenging illumination conditions.
Collaborating with AWE Company (external link) , we developed a large-scale collaborative Augmented Reality solution that can localize multiple users in a challenging outdoor environment in real-time. The project started with a hardware tracking solution and eventually migrated to a novel visual-based tracking algorithm, and deployed at the Fort York National Historic Site (external link) in Toronto for the 200 year anniversary of the war of 1812. The project garnered significant media attention, with coverage in the Toronto Star (external link) and the Space Channel (external link) .
