Deep learning in radiology – hype, reality or the future? Lessons from a collaborative journal club intersecting radiologists and scientists
Judy W. Gichoya, M.D., M.S.
Friday, February 9 at 12:30 p.m. in IT 252
Deep learning continues to have a large adoption in big technology companies for example labeling dogs and cats. The same techniques are now increasingly being applied to medical imaging with numerous publications claiming super human performance in diagnostic performance. This talk will explore some of the papers recently published claiming super human performance, their clinical relevance and data quality issues. At the end of the talk, you will have knowledge of how to approach deep learning in medical imaging, while understanding limitations that currently exist.
About Judy Gichoya
Judy Gichoya is final year radiology resident at the Indiana School of Medicine. She came to the US as a clinical informatics fellow at the Regenstrief Institute in 2012, also enrolled as an informatics student at IUPUI then. Her early work was around public health informatics, where Judy analyzed patterns of health behavior among patients who have notifiable conditions in the Indiana Health Information Exchange.
Judy’s passion is on open source, global health, STEM diversity and most importantly, centering her work around the consultation room – specifically areas where doctors and other health care providers deliver care. She sees an opportunity to improve health care using technology. Her informatics expertise is around innovation, working to build products that improve the way health care is delivered to change lives. Her current projects include developing an open source solution radiology information solution for low resource settings, radiology informatics including machine learning and deep learning and building social enterprises for health. Her skills are in programming, project evangelism, mentorship, academic and grant writing, and clinical skills.