A generative AI-based video interpolation model called Angio-FILM has been developed to produce smooth coronary angiography images at half the standard frame rate, cutting radiation exposure during the procedure by more than half. Developed by researchers at Seoul National University Bundang Hospital, the system records at about 7.5 frames per second and uses AI to generate realistic intermediate frames.
Published in npj Digital Medicine, the study is significant because physical equipment improvements to reduce radiation dose have already reached their limits. The AI approach offers a breakthrough that could protect both patients and medical staff from cumulative radiation exposure during coronary angiography, a procedure essential for diagnosing and treating coronary artery disease. Clinical reliability has been confirmed, paving the way for field introduction.
