The Center for Intelligent Imaging (CI2) brings together the world's leading experts in Radiology & Biomedical Imaging, the state-of-the-art in computational infrastructure, innovative image analytics and artificial intelligence tools, and the rich clinical data from the world's leading medical center to achieve one goal: to harness the power of imaging to improve patient care.

The UCSF Department of Radiology and Biomedical Imaging, with its extensive annotated image archives, massive image databases developed out of research studies and clinical trials, and domain knowledge at all organ and disease levels, is positioned to lead in the discovery, innovation, and translation of intelligent imaging. With this in mind we established the Center for Intelligent Imaging, an institutional resource focused on applications of artificial intelligence and image analysis tools in medical imaging.

The Center is comprised of clinical radiologists, imaging scientists, engineers, machine learning scientists, data engineers, clinicians, post-doctoral fellows, and students collaborating to develop and deploy artificial intelligence that will solve critical clinical problems by advancing the way in which healthcare professionals are able to utilize and deliver imaging.

What is Intelligent Imaging at UCSF?

The Center for Intelligent Imaging provides the internal UCSF community and external academic and industry partners a unique resource in which to discover, innovate, and translate artificial intelligence tools to improve patient care. The Center is unique in its dedicated focus on medical imaging by leveraging the power of data.

The Center for Intelligent Imaging aims to develop strategies that leverage computational tools to improve image acquisition, workflow efficiency, quantitative imaging, and automated diagnosis at 4 key points:

Robust Image Acquisition & Analysis

There are countless sources of variability in medical images. One major goal of the Center is to improve the value of Radiology by reducing systemic variability. By standardizing acquisition for each patient, it will be possible to drive down this variability and achieve faster, safer, and higher-quality images in patients undergoing imaging examinations. The ability to robustly acquire data, reduce noise, tailor parameters of exams for patients, and derive quantitative metrics from imaging data are essential to achieve consistent patient outcomes.

Quantitative Imaging

Quantitative Analysis is key to improved diagnosis and prediction. Tools under development include organ and tissue segmentation, automated volumetry and morphological quantification, and disease visualization. Together these tools provide a computational "cockpit" for objective disease characterization to expedite diagnosis, improve accuracy and enhance therapeutic decisions. Using time series of images also enables modeling of natural history, disease progression, and therapeutic response to optimize patient care.

Translation to Clinical Practice

By providing imaging databases, annotation models, and visualization tools, we can nimbly implement and evaluate new techniques for Radiologists and for referring clinical services.

Enterprise Integration

Offering domain expertise to integrate imaging across the enterprise, we can deliver ready access to clinical data for big data analytics in epidemiology and population health studies.