Machine Learning For Medical Imaging (ML4MI)
The Machine Learning For Medical Imaging (ML4MI) Initiative exists to foster interdisciplinary collaboration between machine learning (ML) experts and medical imaging researchers at the University of Wisconsin, in order to develop and apply state-of-the-art ML solutions to challenging problems in medical imaging. The ML4MI initiative, co-organized by Diego Hernando, Po-Ling Loh, and Varun Jog, includes monthly seminars, a one-day workshop, and a pilot grant program. For more information and updates, see the full ML4MI website.
The first ISMRM sponsored Workshop on Fat-Water Separation, co-organized by Diego Hernando and Scott Reeder in 2012, brought together experts in imaging and medicine to discuss developments in imaging body and organ fat deposition. This workshop included historical and clinical perspectives, as well as updates on advanced quantitative biomarkers and development of applications throughout the body. As part of this workshop, the organizers, led by Diego Hernando and Harry Hu, compiled a Fat-Water Toolbox that included multiple recent (at the time) algorithms and datasets for fat-water separation (ISMRM login required). Over the past few years, this toolbox has facilitated the dissemination of algorithms for applied MRI research, and the development of new algorithms, eg. as part of the 2012 ISMRM Reconstruction Challenge.
Sometimes I get questions about how to choose echo times for fat-water separation or PDFF/R2* quantification. Here is a Jupyter Notebook with some interactive simulations and calculations: