John Hayes
Position title: Research Assistant
Email: jwhayes@wisc.edu
Phone: (616) 217-7245
Address:
1111 Highland Avenue, 1005 WIMR 1122-R1, Madison, WI, 53705
Education
- B.S., Nuclear Engineering and Radiological Sciences, University of Michigan (2009)
- M.S., Nuclear Engineering, University of Michigan (2010)
- M.S. Medical Physics, University of Wisconsin (2016)
Research Interests
My research focuses on developing enabling technologies for low dose computed tomography (CT) as well as characterizing their imaging performance. Software technologies include image reconstruction algorithms such as model based iterative reconstruction (MBIR), as well as data correction methods such as low signal correction (LSC). Imaging performance is assessed based on these important image quality features:
- Noise streaks
- Noise magnitude
- Spatial resolution
- CT number accuracy
- Task-based detectability (human and model observers)
Ideally, we seek methods that preserve the well-known imaging performance properties predicted using linear systems theory such as spatial resolution that is independent of dose and contrast, and noise variance that is inversely proportional to dose. CT number bias at low dose is an example of an issue that was observed empirically for a long time but lacked a solid theoretical framework. Recently, our group provided this theoretical framework, and we continue to work on methods that will reduce bias. Finally, detectability models and studies with human observers integrate each of these image quality features with specific imaging tasks and represent state-of-the-art methods to determine the applicability of an image to be used in the clinic.