A ‘before’ and ‘after’ photo of a plate of french fries, an open-faced hamburger with cheese and bacon on a bun and a green salad whose carbs are being measured by AI.

Counting carbs with AI for real-time glucose monitoring

Diabetes patients have to monitor their diet’s glucose levels closely to avoid serious health complications. While some tools exist to help manage this challenging disease, one that accurately pre-evaluates diabetes patients’ meals and allows for on-the-spot portion adjustments is lacking.

To fill this industry gap and improve the lives of diabetes patients, three Toronto Metropolitan University (TMU) engineering alumni connected with TMU biomedical engineering professor Naimul Khan to develop machine-learning algorithms capable of analyzing 2D food images for 3D depth in real-time. This innovation allows users to snap a photo of their meal and have the carbohydrates counted while they wait, allowing them to adjust their portions or food choices to maintain ideal glucose levels.

Headshots of TMU professor Naimuil Khan and alumni Liam Bell, Osama Muhammad and Muhammed Ashad Khan
Clockwise from top left: Biomedical engineering professor Naimul Khan and alumni Liam Bell (biomedical), Osama Muhammad (mechanical), and Muhammed Ashad Khan (electrical) worked together through Mitacs to improve glucose self-monitoring

Alumni Liam Bell (biomedical), Osama Muhammad (mechanical) and Muhammed Ashad Khan (electrical) are using these algorithms to further develop their smartphone app and accompanying wearable device, Glucose Vision. This technology has the potential to significantly reduce future health issues and the cost burden of diabetes on the Canadian health care system.

Funding for this project provided by Mitacs. To learn more about how Mitacs supports groundbreaking research and innovation, visit the Mitacs website.