A NOVEL METHOD COMBINING MOTION CAPTURE, EYE TRACKING AND 3D MODELING TO CHARACTERIZE HUMAN VISUOSPATIAL BEHAVIOR IN COMPLEX REAL ENVIRONMENTS

Studying human behavior in natural environments poses significant challenges for experimenters. Most methods either take place in laboratory conditions, or must trade some freedom in the post experiment analysis for real world conditions. To address these issues, studies in virtual reality have been proposed. These studies open the door for a lot of experimentations, but fail to address completely the ecological experiment challenge. Some studies have also been proposed by wearing eye tracking and motion capture devices, allowing experimenters to propose completely ecological studies, but therefore losing analysis power offered by VR. What could be done to get the best of both worlds ? We propose a new method that aims to bridge the gap between the two. It consists in merging mobile eye tracking, motion capture technology and a 3D reconstruction of the environment. Using our method, we could accurately and automatically detect fixations on regions of interest while our participants performed a spatial navigation task in a complex environment. We managed to obtain a gaze vector accuracy of 1 meter and 9 degrees for targets as far as 10 meters. This allows us to gather data precise enough for most behavioral experiments, while keeping both perfect ecological conditions from in site studies and experimenter freedom in data manipulation from VR and laboratory settings. To illustrate the usefulness of our method, we present data from a spatial navigation experiment where young and old participants had to follow a sign in the environment.