FRDF: Face Recognition Using Fusion of DTCWT and FFT Features☆

Abstract Face recognition is a physiological Biometric trait widely used for personal authentication. In this paper we have proposed a technique for face recognition using Fusion of Dual Tree Complex Wavelet Transform (DTCWT) and Fast Fourier Transform (FFT) features. The Five Level DTCWT and FFT are applied on the pre-processed face image of size 128 × 512. The Five Level DTCWTfeatures are arranged in a single column vector of size 384 × 1. The absolute values of FFT features are computed and arranged in column vector of size 65, 536 × 1. The DTCWT features are fused with dominant absolute FFT values using arithmetic addition to generate a final set of features. The test image features are compared with database features using Euclidean distance to identify aperson. The face recognition is performed for different database such as ORL, JAFFE, L-SPACEK and CMU-PIE having different illumination and pose conditions. It is observed that the performance parameters False Acceptance Rate (FAR), False Rejection Rate (FRR) and True Success Rate (TSR) of proposed method FRDF: Face Recognition using Fusion of DTCWT and FFT are better compared to existing state of the art methods.

[1]  K.B. Raja,et al.  Performance Evaluation of Face Recognition Based on DWT and DT-CWT Using Multi-matching Classifiers , 2011, 2011 International Conference on Computational Intelligence and Communication Networks.

[2]  Ali Moeini,et al.  Real-World and Rapid Face Recognition Toward Pose and Expression Variations via Feature Library Matrix , 2015, IEEE Transactions on Information Forensics and Security.

[3]  Jufu Feng,et al.  Downsampling sparse representation and discriminant information aided occluded face recognition , 2014, Science China Information Sciences.

[4]  Liu Song,et al.  Face recognition based on 2DPCA and DWT , 2011, Proceedings of 2011 Cross Strait Quad-Regional Radio Science and Wireless Technology Conference.

[5]  Chengyi Xiong,et al.  A robust face recognition method using multiple features fusion and linear regression , 2014, Wuhan University Journal of Natural Sciences.

[6]  Jian Yang,et al.  Nearest orthogonal matrix representation for face recognition , 2015, Neurocomputing.

[7]  N. Kingsbury.  Complex Wavelets for Shift Invariant Analysis and Filtering of Signals , 2001 .

[8]  Nick G. Kingsbury,et al.  A dual-tree complex wavelet transform with improved orthogonality and symmetry properties , 2000, Proceedings 2000 International Conference on Image Processing (Cat. No.00CH37101).

[9]  Arif Mahmood,et al.  Hyperspectral Face Recognition With Spatiospectral Information Fusion and PLS Regression , 2015, IEEE Transactions on Image Processing.

[10]  K. B. Raja,et al.  Face recognition using DT-CWT and LBP features , 2012, 2012 International Conference on Computing, Communication and Applications.

[11]  L. M. Patnaik,et al.  Face recognition using transform domain texture features , 2013, Other Conferences.

[12]  Changyin Sun,et al.  Face recognition using DT-CWT feature-based 2DIFDA , 2011, Proceedings of the 30th Chinese Control Conference.

[13]  Lalit M. Patnaik,et al.  An OLBP Based Transform Domain FaceRecognition , 2014 .