Boosting radial strings for 3D face recognition with expressions and occlusions

Loading...
Thumbnail Image
File version

Accepted Manuscript (AM)

Author(s)
Yu, Xun
Gao, Yongsheng
Zhou, Jun
Griffith University Author(s)
Primary Supervisor
Other Supervisors
Editor(s)

Liew, AWC

Lovell, B

Fookes, C

Zhou, J

Gao, Y

Blumenstein, M

Wang, Z

Date
2016
Size
File type(s)
Location

Gold Coast, AUSTRALIA

License
Abstract

In this paper, we present a new radial string representation and matching approach for 3D face recognition under expression variations and partial occlusions. The radial strings are an indexed collection of strings emanating from the nose tip of a face scan. The matching between two radial strings is conducted through a dynamic programming process, in which a partial matching mechanism is established to effectively find those un-occluded substrings. Moreover, the most discriminative and stable radial strings are selected optimally by the well-known AdaBoost algorithm to achieve a composite classifier for 3D face recognition under facial expression changes. Experimental results on the GavabDB and the Bosphorus databases show that the proposed approach achieves promising results for human face recognition with expressions and occlusions.

Journal Title
Conference Title

2016 INTERNATIONAL CONFERENCE ON DIGITAL IMAGE COMPUTING: TECHNIQUES AND APPLICATIONS (DICTA)

Book Title
Edition
Volume
Issue
Thesis Type
Degree Program
School
Publisher link
Patent number
Funder(s)
Grant identifier(s)
Rights Statement
Rights Statement

© 2016 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.

Item Access Status
Note
Access the data
Related item(s)
Subject

Pattern recognition

Computer vision

Persistent link to this record
Citation