Paper 13305-56
Paper 13305-56
Super-resolution optical coherence tomography using a physics-informed diffusion model
29 January 2025 • 10:30 AM - 10:45 AM PST | Moscone South, Room 203 (Level 2)
Abstract
This study introduces a novel super-resolution (SR) and noise suppression method in optical coherence tomography (OCT) images using diffusion models (DM). To that end, a physics-informed DM is developed to learn an inverse function for reversing the degradations in OCT images due to defocus and digital sampling. The proposed method resulted in resolution enhancement and speckle noise removal in OCT images of various sample types including the human cornea, acquired using a Line-Field OCT (LF-OCT) system. By delivering noise-free and sharpened images at high digital resolutions, the proposed method can potentially facilitate tasks such as retrieving complex point spread functions (PSFs), thereby enabling more precise aberration correction in OCT images.
Presenter
Univ. of Waterloo (Canada)
Nima Abbasi obtained a B.Sc. degree in Electrical Engineering from Sharif University of Technology in 2017, followed by a MASc. degree in Electrical and Computer Engineering from the University of Waterloo in 2019. After working in the field of medical imaging for two years, he started a Ph.D. in Biomedical Engineering at the University of Waterloo in 2022. He is now pursuing Ph.D. within the departments of Systems Design Engineering and Physics at the University of Waterloo, under the co-supervision of Dr. Bizheva and Dr. Wong. His research focuses on ultra-high resolution ultra-fast ophthalmic imaging using OCT.