Skip to content

SDM-based means of gradient for eye center localization

Research output: Chapter in Book/Report/Conference proceedingConference contribution

For eye gaze estimation and eye tracking, localizing eye center is a crucial requirement. This task is challenging work because of the significant variability of eye appearance in illumination, shape, color and viewing angles. In this paper, we improve the performance of means of gradient method in low resolution images, which could locate the eye center more accurately. The proposed method applies Supervised Descent Method (SDM), which has remarkable achievement in the field of face align-ment, to improve the traditional means of gradient method in localizing eye center. We extensively evaluate our method on BioID database which is very challenging and realistic for eye center localization. Moreover, we have compared our method with existing state of the art methods and the results of the experiment confirm that the proposed method is an attractive alternative for eye center localization.
Original languageEnglish
Title of host publication2018 IEEE 16th Intl Conf on Dependable, Autonomic and Secure Computing, 16th Intl Conf on Pervasive Intelligence and Computing, 4th Intl Conf on Big Data Intelligence and Computing and Cyber Science and Technology Congress (DASC/PiCom/DataCom/CyberSciTech)
Number of pages6
ISBN (Electronic)978-1-5386-7518-2
ISBN (Print)978-1-5386-7519-9
Publication statusPublished - 29 Oct 2018
Event16th IEEE International Conference on Pervasive Intelligence -
Duration: 12 Aug 201815 Aug 2018


Conference16th IEEE International Conference on Pervasive Intelligence


  • SDM-based means of gradient for eye center localization

    Rights statement: © © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

    Accepted author manuscript (Post-print), 568 KB, PDF document

Related information

Relations Get citation (various referencing formats)

ID: 10937803