Gaze estimation aims to obtain the direction where you are looking at. It has been actively studied in the past three decades due to its importance in various fields, e.g. human-machine interfaces, human behavior analysis, marketing research, driver safety analysis, cognitive science, and psychology. The recent ten years have seen great performance improvements in gaze estimation methods, however, there is still a huge gap in applying these methods directly into daily webcam-based environments due to the challenges caused by low resolutions, large head movements, eye appearance variations.
This project will focus on developing accurate deep learning-based gaze estimation methods to bridge this gap. Different deep learning models like Convolutional Neural Networks, Generative Adversarial Networks will be explored to improve the accuracy of gaze estimation in common human interaction scenarios. This project is meaningful in that not only does it offer you an opportunity to learn state-of-the-art deep learning methods, but it also enables you to develop practical gaze related applications.