E. Wood, T. Baltrušaitis, L.-P. Morency, P. Robinson, A. Bulling, Learning an Appearance-based Gaze Estimator from one million Synthesised Images, In Proceedings of the Symposium on Eye Tracking Research & Applications (ETRA), 2016
The eyes and their movements convey our attention, indicate our interests, and play a key role in communicating social and emotional information. Estimating eye gaze is, therefore, a significant problem for computer vision, with applications in social signal analysis, health behavior informatics, and gaze-based interfaces. Estimating gaze remotely under unconstrained lighting conditions and dramatic head-pose variations without specialty equipment is a very challenging problem with broad applications.
In our work, we explore two types of methods for tackling the problem: appearance-based and morphable model-based. Appearance-based methods that directly estimate gaze from an eye image have recently improved upon person-independent and device-independent gaze estimation by learning invariances from large amounts of labeled training data. 3D morphable models are a powerful tool as they combine a model of face variation with a model of image formation, allowing pose and illumination invariance. We further explore the use of computer graphics methods to generate training data for appearance and morphable model-based systems.