26 August 2004 Face recognition by using optical correlator with wavelet preprocessing
Author Affiliations +
Proceedings Volume 5566, Optical Security and Safety; (2004) https://doi.org/10.1117/12.577211
Event: Optical Security and Safety, 2003, Warsaw, Poland
Abstract
The method of face recognition by using optical correlator with wavelet preprocessing is presented. The wavelet transform is used to improve the performance of standard Vander Lugt correlator with phase only filter (POF). The influence of various wavelet transforms of images of human faces on the recognition results has been analyzed. The quality of the face recognition process was tested according to two criteria: the peak to correlation energy ratio (PCE), and the discrimination capability (DC). Additionally, proper localization of correlation peak has been controlled. During the preprocessing step a set of three wavelets -- mexican hat, Haar, and Gabor wavelets, with various scales was used. In addition, Gabor wavelets were tested for various orientation angles. During the recognition procedure the input scene and POF are transformed by the same wavelet. We show the results of the computer simulation for a variety of images of human faces: original images without any distortions, noisy images, and images with non-uniform light ilumination. A comparison of results of recognition obtained with and without wavelet preprocessing is given.
© (2004) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Jacek Strzelecki, Katarzyna Chalasinska-Macukow, "Face recognition by using optical correlator with wavelet preprocessing", Proc. SPIE 5566, Optical Security and Safety, (26 August 2004); doi: 10.1117/12.577211; https://doi.org/10.1117/12.577211
PROCEEDINGS
7 PAGES


SHARE
Back to Top