Data Reductionality Technique for Face Recognition | [Proceedings of Pakistan Academy of Sciences • 2011]

Author(s):

1. Muhammad Sharif: Department of Computer Sciences, COMSATS Institute of Information Technology,Wah Cantt.,Pakistan

2. Muhammad Kamran Ayub: Department of Computer Sciences, COMSATS Institute of Information Technology,Wah Cantt.,Pakistan

3. Mudassar Raza: Department of Computer Sciences, COMSATS Institute of Information Technology,Wah Cantt.,Pakistan

4. Sajjad Mohsin: Department of Computer Sciences, COMSATS Institute of Information Technology,Wah Cantt.,Pakistan

Abstract:

The techniques of data reduction to display graphics play a vital role in saving cost and computational complexities. Principal Components Analysis (PCA) is such a well known technique for data compression. In this paper, a comparison of PCA-based face recognition technique and the proposed hybrid technique has been given. In the proposed technique, data compression is achieved through the combination of both Discrete Cosine Transform (DCT) and PCA. It is, in fact, an effort to get the reduced dimensions of image without losing imperative information. The idea of proposed work is based on getting pure co-efficients from the features of a face image. The performance of proposed algorithm is tested and the results obtained show significant improvement over the traditional PCA technique.

Page(s): 233-238

DOI: DOI not available

Published: Journal: Proceedings of Pakistan Academy of Sciences, Volume: 48, Issue: 4, Year: 2011

Keywords:

face , dimensionality , DCT , Data reduction , recognition

References:

References are not available for this document.

Citations

Citations are not available for this document.

Citations

Downloads

Views