[Abstract]

Current clinical workflows primarily rely on 2D imaging, such as cardiac cine magnetic resonance (CMR), for cardiac examinations. These modalities provide only 2D views, which severely limits our understanding of the true 3D structure and dynamic motion of the heart. At the same time, the high privacy of medical data lead to data scarcity, further hindering progress in cardiac research. This talk will present recent advances by our team in AI-powered cardiac modelling, covering several works accepted by IEEE Transactions on Medical Imaging and MICCAI. The content includes 3D digital heart reconstruction from multi-view 2D CMR images, 3D cardiac motion tracking methods, and the use of diffusion models to generate high-quality dynamic medical images (e.g., echocardiography videos). These methods demonstrate significant advantages in modelling accuracy, data generation capability, and clinical applicability, offering new directions for the application of AI in cardiovascular image analysis.

 

[Biography]

Dr. Qingjie Meng is an Assistant Professor in the School of Computer Science at the University of Birmingham. She is also an Honorary Research Associate at Imperial College London. She received her PhD from Imperial College, working on AI for medical image analysis. Her research focus is medical AI, with primary interests including the digital heart, medical data generation, transfer learning, and representation learning. She has published extensively in top-tier journals and conferences such as IEEE Transactions on Medical Imaging (TMI), IEEE Transactions on Image Processing (TIP), CVPR, and MICCAI. She serves as an Area Chair for MICCAI and has organized multiple workshops at international conferences including ECCV, ICCV, and MICCAI.