Abstract

In this research an innovative fully automated 3D face compression and recognition system is presented. Several novelties are introduced to make the system performance robust and efficient. These novelties include: First, an automatic pose correction and normalization process by using curvature analysis for nose tip detection and iterative closest point (ICP) image registration. Second, the use of spherical based wavelet coefficients for efficient representation of the 3D face. The spherical wavelet transformation is used to decompose the face image into multi-resolution sub images characterizing the underlying functions in a local fashion in both spacial and frequency domains. Two representation features based on spherical wavelet parameterization of the face image were proposed for the 3D face compression and recognition. Principle component analysis (PCA) is used to project to a low resolution sub-band. To evaluate the performance of the proposed approach, experiments were performed on the GAVAB face database. Experimental results show that the spherical wavelet coefficients yield excellent compression capabilities with minimal set of features. Haar wavelet coefficients extracted from the face geometry image was found to generate good recognition results that outperform other methods working on the GAVAB database.

Highlights

  • Representing and recognizing objects are two of the key goals of computer vision systems [1,2,3,4,5]

  • iterative closest point (ICP) attempts to iteratively refine the transformation M consisting of a rotation R, and translation T, which minimizes the average distance between corresponding closest pairs of corresponding points on the two meshes

  • In this paper an innovative approach for 3D face compression and recognition based on spherical wavelet parameterization was proposed

Read more

Summary

INTRODUCTION

Representing and recognizing objects are two of the key goals of computer vision systems [1,2,3,4,5]. Pose variations can be accounted for by complete transformations (rotation and translations) between different 3D images computed in the 3D space. This efficiently removes the transformation out of the image plane, which is very difficult in 2D face recognition. The system utilizes discriminative spherical wavelet coefficients which are robust to expression and pose variations to efficiently represent the face image with a small set of features. The ICP algorithm is used to align the face image and to normalize the effect of face poses and position variations This registration process typically applies rigid transformations such as translation and rotation on the 3D faces in order to align them.

RELATEDWORK
Image smoothing
Nose tip detection
Face Registration
SPHERICAL WAVELET PARAMETRIZATION
Spherical Parameterization
Geometry Image
Wavelet Transform Haar Transform
Dimensionality Reduction
EXPERIMENTAL RESULTS
CONCLUSION
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call