KNUSTSpace >
Research Articles >
College of Science >

Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/12821

Title: Recognition of face images under varying head-poses using fft-pca/svd algorithm
Authors: Avuglah, Richard K.
Asiedu, Louis
Nortey, Ezekiel N. N.
Yirenkyi, Favour N.
Keywords: fast Fourier transform
repeated measures design
principal component analysis
singular value decomposition
face recognition algorithm
Issue Date: 2018
Publisher: Pushpa Publishing House, Allahabad, India
Abstract: The human brain has the inborn characteristics to distinguish between faces. The human brain has particular nerve cells for responding to local features of a scene, such as edges, angles, lines or movements. Automated intelligent systems have been developed to mimic this skill inherent in human beings. These systems extract meaningful features from an image, put them into useful representations and classify them. Amidst these achievements is the challenge of recognizing face images under varying constraints. This study assessed the performance of principal component analysis with singular value decomposition using fast Fourier transform for preprocessing (FFT-PCA/SVD) face recognition algorithm under specified angular constraints (4°, 8°, 12°, 16°, 20°, 24°, 28° and 32°). Ten face images from 10 subjects captured under straight-pose (0°) were used for training in the face recognition module. Eighty face images from 10 subjects captured under the specified constraints were used for testing. The study adopted the repeated measures design to ascertain whether statistically significant differences exist in the average recognition distance of the various angular constraints from their straight-pose when the FFT-PCA/SVD is used for recognition. The results of the study revealed that statistically significant differences exist in average recognition between all head-poses except those that are at 4° or less apart. The study also found that FFT-PCA/SVD has a high average recognition rate of 92.5% with corresponding error rate of 7.5%. The study algorithm (FFT-PCA/SVD) recognizes perfectly (100% recognition rate) head-poses that are 24° and below. FFT-PCA/SVD is therefore recommended for recognition of face images under varying headposes.
Description: An article published by Pushpa Publishing House, Allahabad, India and also available at http://dx.doi.org/10.17654/MS103111769
URI: http://hdl.handle.net/123456789/12821
Appears in Collections:College of Science

Files in This Item:

File Description SizeFormat
RecognitionofFaceImages.pdf627.44 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0! DSpace Software Copyright © 2002-2010  Duraspace - Feedback