Eye-gaze tracking system based on particle swarm optimization and BP neural network

In order to enhance the practicability and accuracy of the eye-gaze tracking system, a new type low pixel eye feature point location method is adopted. This method can accurately extract the eye-gaze features, namely iris centre point and canthus points when the image pickup requirements are low. Th...

Full description

Saved in:
Bibliographic Details
Published in2016 12th World Congress on Intelligent Control and Automation (WCICA) pp. 1269 - 1273
Main Authors Liling Yu, Jiangchun Xu, Shengwang Huang
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.06.2016
Subjects
Biological neural networks
Face detection
Iris
Mathematical model
Particle swarm optimization
Training
Online AccessGet full text

Cover

More Information
Summary:In order to enhance the practicability and accuracy of the eye-gaze tracking system, a new type low pixel eye feature point location method is adopted. This method can accurately extract the eye-gaze features, namely iris centre point and canthus points when the image pickup requirements are low. The eye-gaze tracking method based on particle swarm optimization (PSO) BP neural network is raised, to capture pictures of eyes under the same environment, and a regression model where the connection weights and threshold values are optimized by PSO algorithm is built via BP network. This method is free of the inherent defects of BP network. This method requires only a common camera and normal illumination intensity rather than high-standard hardware, which greatly cuts the restrictive requirements for the system hardware and thus enhances the system practicability. The experiment results show that PSO-BP model is of higher robustness and accuracy than BP model, and is of higher recognition rate and can effectively enhances the eye-gaze tracking accuracy.
DOI:10.1109/WCICA.2016.7578296