A novel FNP-pose estimation for three-dimensional face recognition using DPCA under facial expression

June Youn Hwang*, W. L. Woo, S. S. Dlay

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper presents an efficient 3D face recognition algorithm with facial expression. The proposed algorithm describes a novel FNP (Fast Nose Points) -Pose Estimation which is the Triangle-based four point method to estimate the pose of three-dimensional (3-D) face, (i.e. the 3-D shape). Firstly, we find the specific points using Angle comparison Trace. Secondly, from this triangle we calculate weight point of triangle which is used for translation and compensate the rotation of 3D test facial image. Finally, Depth PCA is employed for 3D face recognition where it performs PCA on a 2D x-y axis and takes into account the depth information from 3D Vertices and Face entries. 2D texture information is mapped corresponding to each point at centre of vertices and segment 2-by-2 around this point. The proposed algorithm allows the use of fast iterative algorithm to compute the 3-D facial pose and 3D face recognition that best fits the data. The algorithm has been tested with 3D database and obtained results provide a high level of robustness and accurate recognition.

Original languageEnglish
Title of host publication2007 15th International Conference on Digital Signal Processing, DSP 2007
PublisherIEEE
Pages235-239
Number of pages5
ISBN (Electronic)1424408822
ISBN (Print)1424408814
DOIs
Publication statusPublished - 13 Aug 2007
Event2007 15th International Conference onDigital Signal Processing, DSP 2007 - Wales, United Kingdom
Duration: 1 Jul 20074 Jul 2007

Conference

Conference2007 15th International Conference onDigital Signal Processing, DSP 2007
CountryUnited Kingdom
CityWales
Period1/07/074/07/07

Fingerprint Dive into the research topics of 'A novel FNP-pose estimation for three-dimensional face recognition using DPCA under facial expression'. Together they form a unique fingerprint.

Cite this