Face Recognition for Limited Memory Applications using Eigenphases of MPEG-7 Fourier Feature Descriptors

N. Zaeri, F. Mokhtarian (UK), and A. Cherri (Kuwait)

Keywords

Face recognition, PCA, MPEG7, and limited memory

Abstract

In this paper, a novel method for face recognition is presented that is based on applying principal component analysis (PCA) to the phase spectrum of the Fourier transform of the covariance matrix constructed from the MPEG-7 Fourier Feature Descriptor vectors. It will be shown that the proposed method maximizes the recognition rate when compared to other methods. Moreover, the dramatic reduction in the accompanied calculations of the proposed system makes it an attractive choice for limited memory applications.

Important Links:



Go Back