Gaze point estimation by regression CNN learning restructured eye images
Kazuma YAGI, Daichi TANIGAWA, Yasuhiko Arai, Yasushi Mae · The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec) · 2024
Gaze point is estimated by regression CNN model which is trained by restructured eye images. The restructured eye images are created and trained under different conditions of image resolution and aspect ratio. Experiments are conducted to investigate which restricted eye images are effective in estimating the gaze point on the screen.