Equitable Medical Imaging

Medical imaging allows us to peer into the body to provide diagnostic information or surgical guidance. Ultrasound systems transmit sound that is reflected and detected by a sensor array placed in contact with the skin, while photoacoustic systems transmit light that is absorbed, causing thermal expansion, generating sound that can be detected with the same array. In both cases, the sensed signals are processed using beamformers to display images. Conventional beamformers, however, exclusively rely on signal amplitudes, ignore the impact of light transmission through darker skin tones, or assume uniform properties which overlook naturally occurring intra- and inter-patient variations. I will provide real-world examples of medical imaging inequities that result from conventional beamformer design choices. I will describe innovative techniques to address these inequities by considering spatial correlations rather than signal amplitudes. Specific clinical applications that have the greatest potential to benefit from such an approach include cardiovascular health assessments, breast cancer diagnosis, biopsies, teleoperated robotic surgery, and wearable devices with flexible arrays.

Speaker Bio:

Muyinatu Bell is the John C. Malone Associate Professor of Electrical and Computer Engineering, Biomedical Engineering, and Computer Science at Johns Hopkins University, where she founded and directs the PULSE Lab. She has a doctorate in biomedical engineering from Duke University and a bachelor's in mechanical engineering from MIT. Among her numerous awards and honors, she has been recognized with MIT Technology Review’s Innovator Under 35 Award, SPIE Early Career Achievement Award, IEEE Ultrasonics Early Career Investigator Award, and NSF Alan T. Waterman Award (2024). Prof. Bell is a Fellow of AIMBE, SPIE, and Optica.

For more information, contact Fauzia Ahmad at fauzia.ahmad@temple.edu.