Collecting and Analyzing Eye-tracking Data in Outdoor Environments

Early eye-tracking experiments required participants to stabilize their heads while looking at two-dimensional stimuli in a room with carefully controlled lighting. Many of the field’s standards for collecting and processing data reflect this history of stationary, laboratory-based eyetracking. Over the last two decades, the cost and quality of wearable eye-tracking systems has improved dramatically, allowing researchers to examine tasks that are performed away from a computer and even outside of the laboratory (for reviews, see Hayhoe & Ballard, 2005; Land, 2006; Land & Tatler, 2009). As head-mounted mobile eye-tracking has become more common, new recommendations for research with wearable trackers have emerged (e.g., chapters 5-8 of Duchowski, 2003; head-mounted tracking sections in Holmqvist et al., 2011). Despite the growing interest in mobile eyetracking, our recent work with outdoor scenes has revealed a need to revise methods for calibrating eye and scene videos, detecting ocular features, parsing the eye movement record into fixations, and comparing eye movements across subjects and time. Throughout this paper, we discuss the challenges that arise with mobile eye-tracking (and with outdoor data collection in particular) and present the solutions we have developed. Together, these provide a method for collecting and analyzing eye-tracking data from outdoor environments.