Multimodal Vascular Imaging
Engineering better images
Finding opportunities to teach the laws of physics to a machine-learning algorithm as it learns to create a medical image is a new and exciting research topic.
![BeckmanIndoors2](https://beckman.illinois.edu/images/default-source/beckman-building-images/beckmanindoors2.jpg?sfvrsn=e7429f07_4)
![A PET image of hindlimb ischemia](https://beckman.illinois.edu/images/default-source/research/multimodal-vascular-imaging.png?sfvrsn=b950c4bd_0)
Researchers use standard ultrasonic equipment to gather information about the force of a probe on the skin and the shape and deformation of the tissue below the probe. It’s entered into a unique combination of standard engineering tools, finite-element algorithms, and artificial neural networks.
They work together to learn the mechanical properties of a patient’s tissues. Mechanical property changes are often the first sign of health problems such as infections and cancers.
View people in this group
Group leader
- Michael Insana