Prof. Bell received the NIH Trailblazer Award from the National Institute of Biomedical Imaging and Bioengineering to support our project entitled, A Machine Learning Alternative to Beamforming to Improve Ultrasound Image Quality for Interventional Access to the Kidney. This project is motivated by the clinical challenges surrounding artifacts in ultrasound images, specifically artifacts caused by multipath scattering and acoustic reverberations (which occur when imaging through the abdominal tissue of overweight and obese patients or visualizing metallic surgical tools). There are no existing solutions to eliminate these artifacts based on today’s signal processing techniques. The goal of this project is to step away from conventional signal processing models and instead learn from raw ultrasound channel data examples with state-of-the-art deep learning techniques that differentiate artifacts from true signals to deliver a new class of clearer, easier-to-interpret ultrasound images that we call CNN-Based images. This work will be completed in collaboration with Austin Reiter, PhD and Kelvin Hong, MD.
Two of our pioneering publications in this area include:
- D Allman, A Reiter, MAL Bell, Photoacoustic source detection and reflection artifact removal enabled by deep learning, IEEE Transactions on Medical Imaging, 37(6):1464-1477, 2018 [pdf | datasets | code]
- AA Nair, T Tran, A Reiter, MAL Bell, A deep learning based alternative to beamforming ultrasound images, IEEE International Conference on Acoustics, Speech and Signal Processing, Calgary, Alberta, Canada, April 15-20, 2018 [pdf]
- Additional related publications are featured here
This work has also been featured in the following articles and press releases:
- SWE Magazine (1/13/17): AI’s Forthcoming Transformation of Medicine
- JHU ECE Department News (3/12/18): Faculty Q&A: Muyinatu (Bisi) Bell
- Deep Learning in Healthcare Summit (5/25/17): Interview with Muyinatu Bell
We additionally have a pending patent for these ideas.