AbstractsEngineering

Images showing regions of blood perfusion in human skin are used to diagnose a wide range of medical conditions. These images are generated using expensive, specialized equipment such as Laser Doppler Imagers. This thesis focuses on applying signal processing techniques to generate blood perfusion heat maps from video taken with a digital camera using Eulerian magnification. In this application, Eulerian magnification amplifies color changes at the pulse rate to make the subtle color variation due to blood perfusion visible to the human eye. Using a MATLAB implementation of the Eulerian magnification algorithm, simulated and real videos of human skin are processed to successfully reveal blood perfusion regions in the face and hand. This research shows that, while Eulerian magnification can be used in this application, it is also very susceptible to impairments from motion and camera configuration. Advisors/Committee Members: Buchner, Marc (Committee Chair).