Digital Screen Detection Enabled by Wearable Sensors: Application in ADL Settings
Tommaso Martire, Payam Nazemzadeh, Alberto Sanna, Diana Trojaniello · 2018
Digital Screens (DS) can be considered harmful to the eyes, if used for a prolonged time, leading to such health issues as computer vision syndrome (CVS) and digital eyestrain. A wearable device able to inform the user about the time spent looking at DS could be helpful in order to suggest him the proper break timings and avoid CVS. In this study, a new Digital Screen Detection (DSD) approach using a new generation of head mounted wearable color light sensor has been proposed. Data from five subjects during common activities of daily living (ADL) in different environments (indoor, outdoor, office, meeting room) characterized by the presence of various DSs (Tablet, PC, large screen) have been acquired. The best performing machine learning algorithms already applied in a previous in-lab controlled study have been analyzed. The results showed that, in ADL context, the Random Forest algorithm reached the best accuracy in DSD with the proposed system (79.3%).