Using Fiducial Marker for Analyzing Wearable Eye-Tracker Gaze Data Measured While Cooking

Abstract

The deficits in cognitive functions affect carrying out instrumental activities of daily living. Assistive products oriented to support these activities, e.g., cooking, should predict the user’s behavior. The user’s areas of interest (AOI) calculated from the eye gaze data measured by a wearable eye-tracker can be used for the prediction. We developed a new AOI estimation method suitable for cooking context based on multiple fiducial markers. The evaluation suggested that our method was more feasible than the use of feature extraction and feature matching for AOI estimation on objects in the kitchen, including cooking utensils with their move and overlap, while cooking.