This paper presents an eye gaze estimation system which robust against various users. Our method utilizes an IR camera mounted on glass to allow user's movement. Pupil knowledge such as shape, size, location, and motion are used. This knowledge works based on the knowledge priority. Pupil appearance such as size, color, and shape are used as the first priority. When this step fails, then pupil is estimated based on its location as second priority. When all steps fail, then we estimate pupil based on its motion as the last priority. The aim of this proposed method is to make the system compatible for various user as well as to overcome problem associated with illumination changes and user movement. The proposed system is tested using several users with various race as well as nationality and the experiment result are compared to the well-known adaptive threshold method and template matching method. The proposed method shows good performance, robustness, accuracy and stability against illumination changes without any prior calibration.

Original languageEnglish
Title of host publicationComputational Science and Its Applications - ICCSA 2010 - International Conference, Proceedings
PublisherSpringer Verlag
Number of pages15
EditionPART 2
ISBN (Print)3642121640, 9783642121647
Publication statusPublished - 2010
Event2010 International Conference on Computational Science and Its Applications, ICCSA 2010 - Fukuoka, Japan
Duration: 23 Mar 201026 Mar 2010

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 2
Volume6017 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference2010 International Conference on Computational Science and Its Applications, ICCSA 2010


  • Eye detection
  • Gaze
  • Pupil
  • Pupil knowledge


Dive into the research topics of 'Improvement of gaze estimation robustness using pupil knowledge'. Together they form a unique fingerprint.

Cite this