Regular graphical pattern detection and its applications

Abstract

While there is no existing cure to Alzheimer's disease, early detection and intervention can greatly improve patient prognosis. However, early signals can be very subtle changes in behaviors. Our research aims to understand an individual's behaviors through analyzing their gaze patterns. We do this by introducing eye tracking in addition to traditional pen-and-paper tests that measure cognitive status. Traditionally, fiducial markers are added to assist in locating gaze positions with respect to an object in the real world. However, fiducial markers can introduce a distraction and make the test different from its traditional pen-and-paper version. To enable eye tracking without fiducial markers, we present an algorithm that identifies the graphics within the test, allowing us to locate a subject's gaze on the test form using the test alone. It is a novel approach to detecting features in regular graphical patterns despite occlusions.