Grant: Idea to Innovation
University of Toronto
$14,888.00
Market Assessment: A Business/Market Opportunity Assessment for High-Performance Semi-Weakly Supervised Learning Models for Pulmonary Embolism Diagnosis
FY 2024-25
Summary
Pulmonary embolism (PE) diagnosis using machine learning, particularly on computed tomography pulmonary angiography, has the potential to revolutionize medical imaging by providing timely and accurate detection of life-threatening conditions. Traditional approaches rely on extensive annotations, which are resource-intensive and hinder practical application. Our research team has developed a high-performance, semi-weakly supervised learning model that effectively balances accuracy and efficiency by combining examination-level labels with a limited portion of slice-level annotations. This innovative approach significantly reduces the need for granular annotations while maintaining high diagnostic accuracy, achieving comparable or superior performance with four to forty times fewer labels than state-of-the-art models.
Our model's integration of Grad-CAM tools enhances interpretability by providing visual explanations for predictions, assisting clinicians in identifying areas likely to contain PE. This combination of efficiency, accuracy, and interpretability positions our technology for immediate use in research and pre-clinical settings, with the potential to streamline clinical workflows and reduce human error.
To capitalize on this innovation, we aim to conduct a thorough market assessment to identify key healthcare institutions, evaluate market size and trends, and understand the competitive landscape. This will enable us to commercialize our technology effectively, ensuring its adoption in clinical settings and addressing current challenges in PE detection.
Our team, consisting of experienced engineers, radiologists, and clinicians, is uniquely positioned to advance this technology, backed by our extensive expertise in machine learning, medical imaging, and computer vision. With a solid intellectual property position and a clear strategy for market entry, we are poised to make significant contributions to the field of computational radiology.
Program
Idea to Innovation
Recipient
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data
As published
Application id
598135-2024
Name nom
Sejdic, Ervin ES
Department département
Electrical & Computer Engineering, Edward S. Rogers Sr. Department of
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
14888
Program id
I2IPJ
Program name en
Idea to Innovation
Program name fr
De l'idée à l'innovation
Group en
RESEARCH PARTNERSHIPS
Group fr
PARTENARIATS DE RECHERCHE
Committee code
96
Committee name en
RPP Internal Decision Cttee
Committee name fr
Cté de décision interne - PPR
Area of application code
1211
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Medical and health sciences
Area of application fr
Sciences médicales et de la santé
Research subject code
2708
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Image processing
Research subject fr
Traitement des images
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
Market Assessment: A Business/Market Opportunity Assessment for High-Performance Semi-Weakly Supervised Learning Models for Pulmonary Embolism Diagnosis
Keywords
machine learning; medical imaging; pulmonary embolism; radiology; semi-weakly supervised learning
Application summary
Pulmonary embolism (PE) diagnosis using machine learning, particularly on computed tomography pulmonary angiography, has the potential to revolutionize medical imaging by providing timely and accurate detection of life-threatening conditions. Traditional approaches rely on extensive annotations, which are resource-intensive and hinder practical application. Our research team has developed a high-performance, semi-weakly supervised learning model that effectively balances accuracy and efficiency by combining examination-level labels with a limited portion of slice-level annotations. This innovative approach significantly reduces the need for granular annotations while maintaining high diagnostic accuracy, achieving comparable or superior performance with four to forty times fewer labels than state-of-the-art models.
Our model's integration of Grad-CAM tools enhances interpretability by providing visual explanations for predictions, assisting clinicians in identifying areas likely to contain PE. This combination of efficiency, accuracy, and interpretability positions our technology for immediate use in research and pre-clinical settings, with the potential to streamline clinical workflows and reduce human error.
To capitalize on this innovation, we aim to conduct a thorough market assessment to identify key healthcare institutions, evaluate market size and trends, and understand the competitive landscape. This will enable us to commercialize our technology effectively, ensuring its adoption in clinical settings and addressing current challenges in PE detection.
Our team, consisting of experienced engineers, radiologists, and clinicians, is uniquely positioned to advance this technology, backed by our extensive expertise in machine learning, medical imaging, and computer vision. With a solid intellectual property position and a clear strategy for market entry, we are poised to make significant contributions to the field of computational radiology.