location: Current position: Home >> Scientific Research >> Paper Publications

Learning a gaze estimator with neighbor selection from large-scale synthetic eye images

Hits:

Indexed by:期刊论文

First Author:Wang, Yafei

Correspondence Author:Fu, XP (reprint author), Dalian Maritime Univ, Informat Sci & Technol Coll, Dalian 116026, Peoples R China.

Co-author:Zhao, Tongtong,Ding, Xueyan,Peng, Jinjia,Bian, Jiming,Fu, Xianping

Date of Publication:2018-01-01

Journal:KNOWLEDGE-BASED SYSTEMS

Included Journals:SCIE、EI

Volume:139

Page Number:41-49

ISSN No.:0950-7051

Key Words:Gaze estimation; Neighbor selection; Learning-by-synthesis; Cross-subject

Abstract:Appearance-based gaze estimation works well in inferring human gaze under real-world condition. But one of the significant limitations in appearance-based methods is the need for huge amounts of training data. Eye image synthesis addresses this problem by generating huge amounts of synthetic eye images with computer graphics. To fully use the large-scale synthetic eye images, a simple-but-effective appearance-based gaze estimation framework with neighbor selection is proposed in this paper. The proposed framework hierarchically fuses multiple k-NN queries (in head pose, pupil center and eye appearance spaces) to choose closest samples with more relevant features. Considering the structure characters of the closet samples, neighbor regression methods then can be applied to predict the gaze directions. Experimental results demonstrate that the representative neighbor regression methods under the proposed framework achieve better performance for within-subject and cross-subject gaze estimation. (C) 2017 Elsevier B.V. All rights reserved.

Pre One:Piezo-phototronic effect enhanced photo-detector based on ZnO nano-arrays/NiO structure

Next One:Kinetics of waterborne fluoropolymers prepared by one-step semi-continuous emulsion polymerization of chlorotrifluoroethylene, vinyl acetate, butyl acrylate andVeova 10