Talk Abstract
AI-powered segmentation of hip and knee bony anatomy has become indispensable in orthopedics, revolutionizing both pre-operative planning and post-operative assessment. Despite the remarkable performance made by deep learning computer vision algorithms in medical image segmentation, a critical concern remains largely unexplored, and that is how to uncover and address potential biases inherent within these AI-powered models. The present contribution undertakes a comprehensive re-examination of deep learning-driven hip and knee bony anatomy segmentation employing plain radiographs, with a specific focus on revealing discernible gender and racial biases. Through an extensive evaluation, this work offers insights into underlying causes of biases, presenting targeted mitigation strategies tailored to alleviate gender and racial biases, thereby engendering automatic segmentation results that are fair, impartial, and safe in the context of AI. As AI-powered medical imaging continues to advance, the lessons gleaned from this study will furnish invaluable intuitions guiding the development of deep learning models toward a more inclusive, ethically sound, equitable, and unbiased healthcare landscape.
Biography
Ahmad P. Tafti is an Assistant Professor of Health Informatics in the Department of Health Information Management within the School of Health and Rehabilitation Sciences at the University of Pittsburgh, with a secondary appointment in the Intelligent Systems Program (ISP), at the School of Computing and Information. He is also leading the Pitt HexAI Research Laboratory, conducting Health and Explainable AI Podcast series. Furthermore, Ahmad is a Fellow of the American Medical Informatics Association, and affiliated with the Center for AI Innovation in Medical Imaging (CAIIMI), also serving as an Associate Member at UPMC Hillman Cancer Center, plus as the Vice Chair of IEEE Computer Society at Pittsburgh. He earned his BS, MS, and PhD all in Computer Science, with a main focus on fundamental and applied AI in healthcare. Ahmad P. Tafti is the 2021 SiiM Imaging Informatics Innovator awardee, Oracle for Research Project awardee, Oracle Eureka Excellence Awardee, Mayo Clinic Transform the Practice awardee, and GE Healthcare Honorable Mention awardee. To date, he has authored 45+ peer-reviewed publications, organizing numerous workshops on intelligent health systems and has served on the program committee of 15+ conferences, symposiums, and journals in AI and health data science.
Location
Sennott Square Building, Room 5317
Date
Friday, September 29 at 2:00 p.m. to 3:15 p.m.