Vision-based gaze estimation improves gaze prediction accuracy in educational environments, suggesting a cost-effective alternative to traditional methods.
This poster illustrates the implementation of a vision-based gaze estimation system related to the method presented in the paper cited below, titled What are you Looking at? A Preliminary Vision-Based Gaze Estimation Algorithm for Humanoid Robots in Educational Environments. The system predicts where a person is looking using images captured from a standard webcam, with the aim of exploring a low-cost, non-intrusive and accessible alternative to expensive traditional eye-tracking devices. The algorithm extracts head posture (yaw, pitch) and eye posture (iris-based descriptors) from facial landmarks and then applies a linear gaze estimation model to estimate the coordinates of the user's screen. As indicated in the study, combining head and eye posture descriptors can improve the precision and accuracy of gaze prediction compared to using head posture alone.
No takes yet. Share an insight, caveat, or question.
Rodríguez-Rubio et al. (2024) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: