Archives


Rate this Article: (0) Votes

Research Article

Year : 2015 | Volume: 1 | Issue: 1 | Pages: 34-41

Spectral Graph Wavelet Theory with Statistical Features for Face Recognition

S Shankari Devi1*, Brindha.2

http://dx.doi.org/doi:10.18831/djcse.in/12015011005

Corresponding author

S Shankari Devi*

Lord Jegannath College of Engineering, Kanyakumari, Tamil Nadu, India.

  • 1. Lord Jegannath College of Engineering, Kanyakumari, Tamil Nadu, India.

Received on: 2015/09/16

Revised on: 2015/10/10

Accepted on: 2015/10/12

Published on: 2015/10/12

  • Spectral Graph Wavelet Theory with Statistical Features for Face Recognition, S Shankari Devi, Brindha., 2015/10/12, Journal of Excellence in Computer Science and Engineering, 1(1), 34-41, http://dx.doi.org/10.18831/djcse.in/12015011005.

    Published on: 2015/10/12

Abstract

Spectral Graph Wavelet Theory(SGWT) is one of the modern techniques in imaging science and is used for face recognition in this project.Spectral Graph Wavelet Theoryimplements an efficient approach for automatic face recognition.Like wavelet transform SGWT is defined atthe  arbitrary finite weighted graph vertices and the transform functions are defined on the  weighted graph vertices.At first the SGWT decay the given face image.The obtained sub-band energies are combined together and the corresponding image is considered as feature vector.Using nearest neighbor classifier,face images in the ORL database is analyzed in the proposed system.The input face image hasposevariations,and also variation in expression and face details,which is used in this study.Thus the proposed system based on Spectral Graph Wavelet Theory produced better result and achieved 96% of recognition accuracy.

Keywords

Robust Face Recognition, Multipartition, Representation of Spectral Graph, Wavelet Transform and Linear Regression Classification.