The mission of MIMI is to develop and validate new techniques to improve the acquisition, interpretation, and ultimately, value of medical imaging technologies. Research in this group combines new machine learning techniques with classical engineering and physics knowledge, with a particular emphasis on solving problems where only limited labeled datasets are available. We explore techniques revolving around semi/self/un-supervised learning, representation learning, multi-modality learning, and model explainability for image analysis and inverse problems. Projects in the group have a focus on improving the efficacy and value of healthcare, and subsequently, entail collaborations amongst faculty and students with technical and clinical expertise. A major principle of the MIMI group is to enable open-sourcing of algorithms and datasets to the broader research community.

TBA

Our group is a part of the Integrative Biomedical Imaging Informatics (IBIIS) and the Precision Health and Integrated Diagnostics (PHIND) divisions in the Department of Radiology at the Stanford School of Medicine. We actively collaborate with the Radiological Sciences Laboratory (RSL), multiple Radiology clinical divisions, the Stanford Center for Artificial Intelligence in Medicine and Imaging (AIMI), Preventive Cardiology, and the Wu-Tsai Human Performance Alliance.