Abstract
Remote eye tracking technology has suffered an increasing growth in recent years due to its applicability in many research areas. In this paper, a video-oculography method based on convolutional neural networks (CNNs) for pupil center detection over webcam images is proposed. As the first contribution of this work and in order to train the model, a pupil center manual labeling procedure of a facial landmark dataset has been performed. The model has been tested over both real and synthetic databases and outperforms state-of-the-art methods, achieving pupil center estimation errors below the size of a constricted pupil in more than 95% of the images, while reducing computing time by a 8 factor. Results show the importance of use high quality training data and well-known architectures to achieve an outstanding performance.
Highlights
Eye tracking technology appeared on the 20th century with the purpose to detect eye position and to follow eye movements
infrared oculography (IRO) methods are based on an infrared emitter that radiates certain amount of light which is reflected in the eye and detected by an infrared detector
Our method achieves an accuracy of 96.68% in the most challenging threshold value, i.e., emax ≤ 0.025, when multi-task cascaded framework based on CNN (MTCNN) face detector is used to create the eye bounding box and an accuracy of 98.46% when the eye bounding box is created using ground-truth landmarks
Summary
Eye tracking technology appeared on the 20th century with the purpose to detect eye position and to follow eye movements. As in many other areas, one of the goals of eye tracking techniques is to be less invasive. Eye tracking methods as the scleral coil or electro-oculography (EOG). Less invasive systems such as infrared oculography (IRO) or video-oculography (VOG) appeared. IRO methods are based on an infrared emitter that radiates certain amount of light which is reflected in the eye and detected by an infrared detector. VOG methods are based on the use of cameras and image processing. Eye position detection and eye movements tracking are performed by detecting specific features related to the shape or appearance of the eye, being the pupil center one of the most important features [11]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
