34th Chaos Communication Congress

»Low Cost Non-Invasive Biomedical Imaging«
2017-12-27, 19:30–20:00, Saal Clarke

An open source biomedical imaging project using electrical impedance tomography. Imagine a world where medical imaging is cheap and accessible for everyone! We'll discuss this current project, how it works, and future directions in medical physics.

Current medical imaging machines such as MRI scanners are large, expensive and very rarely used preventatively as scans are done when symptoms have already occurred. A better healthcare for the future would include affordable high resolution body scans for everyone, which caused no harm to the body and enable us to track changes through machine learning algorithms.

Electrical Impedance Tomography is an electrical current mapping technique enabling the reconstruction of 2D slices of the human body that is both non-invasive and completely safe (non-ionizing). It’s an exciting and active area of research with new techniques coming out all the time to reach higher resolution imaging. The range of applications are huge and include measuring lung volume, muscle and fat mass, gestural recognition based on muscle movement, bladder or stomach fullness, breast and kidney cancer, hemorrhage detection and even monitoring the depth of anesthesia in patients. I’ll talk about the state of research on each of these applications.

Currently there is no readily available platform to enable rapid development and collaboration in this area. Unfortunately this means very few people outside of biomedical engineering R&D have been able to experiment with it. This talk presents a new system in development that enables real-time electrical impedance tomography experimentation. I will present the hardware, python test bench and explanation of how the reconstruction algorithms work, then move to potential future directions and applications of this project.

Democratizing novel sensing technology opens the way to better collaborations and faster innovation to increase human healthspan.