Eye Movement Analysis for Activity Recognition

In this work we investigate eye movement analysis as a new modality\u000A for recognising human activity. We devise 90 different features based\u000A on the main eye movement characteristics: saccades, fixations and\u000A blinks. The features are derived from eye movement data recorded\u000A using a wearable electrooculographic (EOG) system. We describe a\u000A recognition methodology that combines minimum redundancy maximum\u000A relevance feature selection (mRMR) with a support vector machine\u000A (SVM) classifier. We validate the method in an eight participant\u000A study in an office environment using five activity classes: copying\u000A a text, reading a printed paper, taking hand‐written notes, watching\u000A a video and browsing the web. In addition, we include periods with\u000A no specific activity. Using a person‐independent (leave‐one‐out)\u000A training scheme, we obtain an average precision of 76.1% and recall\u000A of 70.5% over all classes and participants. We discuss the most\u000A relevant features and show that eye movement analysis is a rich and\u000A thus promising modality for activity recognition