Search for: All records

Algorithms for the estimation of gaze direction from mobile and video-based eye trackers typically involve tracking a feature of the eye that moves through the eye camera image in a way that covaries with the shifting gaze direction, such as the center or boundaries of the pupil. Tracking these features using traditional computer vision techniques can be difficult due to partial occlusion and environmental reflections. Although recent efforts to use machine learning (ML) for pupil tracking have demonstrated superior results when evaluated using standard measures of segmentation performance, little is known of how these networks may affect the quality of the final gaze estimate. This work provides an objective assessment of the impact of several contemporary ML-based methods for eye feature tracking when the subsequent gaze estimate is produced using either feature-based or model-based methods. Metrics include the accuracy and precision of the gaze estimate, as well as drop-out rate. 

Eye image segmentation is a critical step in eye tracking that has great influence over the final gaze estimate. Segmentation models trained using supervised machine learning can excel at this task, their effectiveness is determined by the degree of overlap between the narrow distributions of image properties defined by the target dataset and highly specific training datasets, of which there are few. Attempts to broaden the distribution of existing eye image datasets through the inclusion of synthetic eye images have found that a model trained on synthetic images will often fail to generalize back to real-world eye images. In remedy, we use dimensionality-reduction techniques to measure the overlap between the target eye images and synthetic training data, and to prune the training dataset in a manner that maximizes distribution overlap. We demonstrate that our methods result in robust, improved performance when tackling the discrepancy between simulation and real-world data samples. 

When humans navigate through complex environments, they coordinate gaze and steering to sample the visual information needed to guide movement. Gaze and steering behavior have been extensively studied in the context of automobile driving along a winding road, leading to accounts of movement along well-defined paths over flat, obstacle-free surfaces. However, humans are also capable of visually guiding self-motion in environments that are cluttered with obstacles and lack an explicit path. An extreme example of such behavior occurs during first-person view drone racing, in which pilots maneuver at high speeds through a dense forest. In this study, we explored the gaze and steering behavior of skilled drone pilots. Subjects guided a simulated quadcopter along a racecourse embedded within a custom-designed forest-like virtual environment. The environment was viewed through a head-mounted display equipped with an eye tracker to record gaze behavior. In two experiments, subjects performed the task in multiple conditions that varied in terms of the presence of obstacles (trees), waypoints (hoops to fly through), and a path to follow. Subjects often looked in the general direction of things that they wanted to steer toward, but gaze fell on nearby objects and surfaces more often than on the actual path or hoops. Nevertheless, subjects were able to perform the task successfully, steering at high speeds while remaining on the path, passing through hoops, and avoiding collisions. In conditions that contained hoops, subjects adapted how they approached the most immediate hoop in anticipation of the position of the subsequent hoop. Taken together, these findings challenge existing models of steering that assume that steering is tightly coupled to where actors look. We consider the study’s broader implications as well as limitations, including the focus on a small sample of highly skilled subjects and inherent noise in measurement of gaze direction. 

The active inference framework (AIF) is a promising new computational framework grounded in contemporary neuroscience that can produce human-like behavior through reward-based learning. In this study, we test the ability for the AIF to capture the role of anticipation in the visual guidance of action in humans through the systematic investigation of a visual-motor task that has been well-explored—that of intercepting a target moving over a ground plane. Previous research demonstrated that humans performing this task resorted to anticipatory changes in speed intended to compensate for semi-predictable changes in target speed later in the approach. To capture this behavior, our proposed “neural” AIF agent uses artificial neural networks to select actions on the basis of a very short term prediction of the information about the task environment that these actions would reveal along with a long-term estimate of the resulting cumulative expected free energy. Systematic variation revealed that anticipatory behavior emerged only when required by limitations on the agent's movement capabilities, and only when the agent was able to estimate accumulated free energy over sufficiently long durations into the future. In addition, we present a novel formulation of the prior mapping function that maps a multi-dimensional world-state to a uni-dimensional distribution of free-energy/reward. Together, these results demonstrate the use of AIF as a plausible model of anticipatory visually guided behavior in humans.