We have developed webcam-based eye gaze tracking system using Attention-Mechanism based deep neural network. Our system reports state of the art performance of 4.09 degree error on MPIIGaze dataset. In this video, we demonstrate that our system can be used in real-time interaction tasks like web browsing and on-screen virtual keyboard typing. Our current system does not require any calibration and can be used on-the-go.
Музыка
Фильмы
ТВ Шоу
Категории видео
