Abstract—This paper addresses the problem of improving face recognition accuracy for local phase quantization (LPQ) descriptor, introduced by Ojansivu et al. in 2008, when recognizing face images under varying conditions. To do this, we propose to apply difference of Gaussians (DoG) for normalizing face images before encoding the obtained images by LPQ and classifying by support vector machines. Experimental results on three databases (the FEI, FERET, and ORL database of faces databases) demonstrated the improvement of the proposed approach from 0.89% to 17.50% compared to LPQ and other descriptors (CS-LBP, LBP, LDP, LTP, and RLBP) and a combination of them with illumination preprocessing methods (DoG, histogram equalization, Gradient faces, self-quotient image, Tan and Triggs, and Weber-face) using the same classification technique. These results indicated that the introduced approach was robust against variations in illumination, pose, expression, occlusion, scale, and age.
Index Terms—Face recognition, difference of Gaussians, local phase quantization, support vector machines.
Chi-Kien Tran, Thanh-Hoa Ngo, Cam-Ngoan Nguyen, and Lan-Anh Nguyen are with Faculty of Information Technology, Hanoi University of Industry, Bac Tu Liem district, Hanoi, Vietnam (e-mail: chikien.tran@haui.edu.vn).
[PDF]
Cite:Chi-Kien Tran, Thanh-Hoa Ngo, Cam-Ngoan Nguyen, and Lan-Anh Nguyen, "SVM-Based Face Recognition through Difference of Gaussians and Local Phase Quantization," International Journal of Computer Theory and Engineering vol. 13, no. 1, pp. 1-8, 2021.
Copyright © 2021 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).
