Learning-by-Synthesis for Appearance-Based 3D Gaze Estimation

Inferring human gaze from low-resolution eye images is still a challenging task despite its practical importance in many application scenarios. This paper presents a learning-by-synthesis approach to accurate image-based gaze estimation that is person- and head pose-independent. Unlike existing appe...

Full description

Saved in:
Bibliographic Details
Published in2014 IEEE Conference on Computer Vision and Pattern Recognition pp. 1821 - 1828
Main Authors Sugano, Yusuke, Matsushita, Yasuyuki, Sato, Yoichi
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2014
Subjects
Cameras
Estimation
Head
Three-dimensional displays
Training
Training data
Vegetation
Online AccessGet full text

Cover

More Information
Summary:Inferring human gaze from low-resolution eye images is still a challenging task despite its practical importance in many application scenarios. This paper presents a learning-by-synthesis approach to accurate image-based gaze estimation that is person- and head pose-independent. Unlike existing appearance-based methods that assume person-specific training data, we use a large amount of cross-subject training data to train a 3D gaze estimator. We collect the largest and fully calibrated multi-view gaze dataset and perform a 3D reconstruction in order to generate dense training data of eye images. By using the synthesized dataset to learn a random regression forest, we show that our method outperforms existing methods that use low-resolution eye images.
ISSN:1063-6919
DOI:10.1109/CVPR.2014.235