Eye tracking system model with easy calibration

Abstract

Calibration is one of the most tedious and often annoying aspects of many eye tracking systems. It normally consists in looking at several marks on a screen in order to collect enough data to modify the parameters of an adjustable model. Unfortunately this step is unavoidable if a competent tracking system is desired. Many efforts have been made to achieve more competent and improved eye tracking systems. Maybe the search for an accurate mathematical model is one of the least researched fields. The lack of a parametric description of the gaze estimation problem makes it difficult to find the most suitable model, and therefore generic expressions in calibration and tracking sessions are employed instead. In other words, a model based on parameters describing the elements involved in the tracking system would provide a stronger basis and robustness. The aim of this work is to build up a mathematical model totally based in realistic variables describing elements taking part in an eye tracking system employing the well known bright pupil technique i.e. user, camera, illumination and screen. The model is said to be defined when the expression relating the point the user is looking at with the extracted features of the image (glint position and center of the pupil) is found. The desired model would have to be simple, realistic, accurate and easy to calibrate.The first step of the work consists in defining the parameters related to each element that present a faithful description of its behavior in the system. For the camera, the intrinsic parameters such as focal distance, aspect ratio and origin of the image plane can be easily found through a camera calibration process. The dimensions of the screen as well as the relative location and orientation related to a main reference system (located on the projection center of the camera) has to be measured. The user and illumination are due to be properly referenced to the same coordinate system. The mathematical description of the eyeball and the parametric definition of its behavior when looking at different points on the screen is one of the most problematic steps of the work. If an accurate description of the eye gaze is desired the differentiation between optical and visual axes of the eye is basic in addition to a deep knowledge of eye kinematics (Listing's and Donder's laws).Trying to perform a reasonable development of the model, in this work the description of the orientation of the optical axis of the eye is presented. This is the mathematical expression relating the optical axis orientation (p x ,p y ) with the information extracted from the captured image i.e. the glint-pupil vector.In order to make a practical development of the model the simplest configuration of the system is proposed, i.e. when the user, screen center and illumination are perfectly aligned on the optical axis of the camera. Once the elements have been properly described in a parametric way, useful geometric relations are established between them to achieve the pursued mathematical expression. ...

Pursuit calibration

“…Schnipke and Todd's account of "trials and tribulations of using an eye-tracking system" [19] might be dismissed as dated but there is hardly any publication of eye-tracked user studies that does not report calibration problems. Villanueva et al, and others in eye tracking research, have described the calibration task as "most tedious" [28].…”

“…A common thrust is to develop advanced geometric models to reduce the number of calibration points that need to be sampled to achieve good accuracy (a generally agreed accuracy target is 1 • of visual angle), for example to only two points [28,16]. Commercial suppliers advertise "childfriendly versions" where only two points are used, and researchers have proposed single-point multi-glint calibration also with infants in mind [6].…”

“…Theoretically fewer points would suffice to establish the input-output mapping, but in practice five or more points are used, spread over the target space. Unfortunately, this is a process that presents users with a task that is of poor usability [14], often experienced as difficult [19], and frequently described as tedious [28,16,33].…”

Wavi

“…Pfeuffer et al [26] improved the accuracy and usability of calibrating an eye tracker by redesigning user feedback and the procedure. Similarly, Villanueva et al [38] proposed new mathematical models to reduce frustration and physical demand. Flatla et al [8] used a gamification approach to create a design framework to disguise calibration tasks as games.…”

Eye tracking: Pupil orientation geometrical modeling

“…Anyway the error is always below the imposed limit. In the work by Villanueva [28] it was demonstrated that it is not the number of points taken for regression but the location of selected marks what affects the competence of the model.…”