This year, John Boby completed his master’s thesis, titled “Cardiac PET Attenuation Correction and Inter-modal Medical Image Translation.” This highly technical project has major real-world implications for cardiac diagnostics. His innovative work was presented at the prestigious IEEE Nuclear Science Symposium, Medical Imaging Conference, and Room Temperature Semiconductor Detectors Symposium for 2025. The event is a platform where top researchers in medical imaging gather to shape the future of health-care technology.
John Boby’s research tackled a complex challenge in hybrid imaging systems, specifically positron emission tomography (PET) / magnetic resonance imaging (MRI) scanners. Unlike PET / computed tomography (CT) systems, PET/MRI scanners struggle with signal attenuation, meaning weakening of the signal intensity. This is an issue that can significantly affect diagnostic accuracy. John Boby’s innovative solution used deep learning to translate MRI images into synthetic CT scans, which could then be used to create more accurate attenuation maps for PET image correction. “We didn’t just apply existing methods,” he explains. “We built a complete pipeline from the ground up, developing a new workflow for multimodal image registration to align MRI and CT scans, then training our model on these registered pairs using a custom loss function optimized for attenuation correction (AC). This approach outperformed the existing AC method at the Royal’s Institute of Mental Health Research (IMHR).”
The model’s potential applications go far beyond academic theory. Institutions like the IMHR and the University of Ottawa Heart Institute — where John Boby worked with experts including his supervisor, Professor Tanya Schmah; his co-supervisor, Professor Robert DeKemp; and collaborator Katie Dinelle — can now use this model in real time. This will improve patient diagnosis and care without the need for extra CT scans.
Yet the path wasn’t always smooth. “Working with 3D medical images is challenging,” John Boby says. “Aligning multimodal scans taken at different times, with even slight shifts in patient posture, was technically difficult. Building a reliable registration workflow was also one of the hardest parts, but I was able to tackle it successfully thanks to the guidance of my supervisor, Professor Tanya Schmah.”
“I want to push the boundaries of what’s possible with AI research while helping organizations apply these advances in fields like health care and finance, where the impact can be transformative.”
John Boby Mesadieu
— MSc graduate John Boby Mesadieu
During his studies, John Boby served as a teaching and research assistant, instructing first- and second-year students in linear algebra, statistics and differential calculus. Through this work, he gained valuable experience in academic instruction.
Now, with his thesis complete and international recognition under his belt, John Boby looks to the future. He plans to work in the industry as a statistician, machine learning engineer or data scientist, with long-term ambitions to contribute to cutting-edge research at premier AI institutions while developing expertise in applied AI consulting. “I want to push the boundaries of what’s possible with AI research while helping organizations apply these advances in fields like health care and finance, where the impact can be transformative,” he says.
His advice for students following a similar path is this: “Question everything, especially the assumptions in your own field. Consistency matters more than perfection. And don’t just network — find mentors, peers and even students who challenge your thinking. The conversations that change your trajectory are rarely the ones you plan.”
With curiosity, determination and a sharp mind for innovation, John Boby is already shaping the future, one image at a time.
Read more: