Parkinson Disease Diagnosis via Automatic Motor Dysfunction Recognition
Natural user interfaces (NUI) such as motion sensors and touch screens are already integrated between humans and environment in diverse settings, and are becoming increasingly critical for every day life. However, precise gestures are often required for proper usage of these interfaces, creating a serious problem for individuals suffering from motor dysfunction. In this study we report the design and implementation of a motion-interpretation layer (MIL) embedded at the user-hardware interface of touch-based NUI. MIL was built using data that was collected from Parkinson’s’s disease patients(PD) and healthy people, and was able to correctly interpret 90% of gestures made by these individuals. In addition to assisting individuals with disabilities to operate NUI, MIL could enable NUI operation during noisy¨tasks such as driving or physical training which can significantly change the way we use the touch screen. Moreover, it can track and analyze motion patterns collected over long usage periods and indicate potential decline in motor function and the early onset of disease.
* This research was carried out towards the Masters degree in Electrical Engineering at Bar-Ilan University, under the supervision of Prof. Yosi Keller and Dr. Ido Bachelet.