Visual Sobriety Test

Abstract: Alcohol and certain drugs impair oculomotor functioning, a phenomenon that leads to jerky eye tracking. This phenomenon is exploited by the Horizontal Gaze Nystagmus (HGN) test, which is performed by law enforcement personnel as a central component of the field sobriety test procedure. The aim of this work is to develop an algorithm to test for HGN.

Our algorithm is composed by three parts. A precycle, in which we extrapolate from the first frame of the video the eye template, the threshold value and the iris radius. These parameters are essential for the second part, the cycle, where we have implemented an eye tracking and a matching method with the goal to get the pupil position for every frame. In the last part, the post-processing, the values acquired are computed to plot the graphics and to spot the eye's jerks.