strDoctor: Indicate Stroke for Elderly through Body Sensing Game

Early diagnosis can significantly reduce the incidence of stroke which is a serious cerebrovascular disease to the elderly. Motivated by the Trail Making Test (TMT) in neuropsychological, a traditional test which uses pen and paper to signal brain damage, we present an intelligent system called strDoctor. Instead of seeing a real doctor, it can guide users to do TMT through body sensing game and give them diagnosis feedback automatically. Depending on a robust palm-pose-adaptive single fingertip tracking method, this paper extracts effective features to construct a stroke indicating model for the body sensing game based TMT. Finally the validity of strDoctor was verified by our experiments.