Hybrid Framework for Robust Multimodal Face Recognition
Both two dimensional principal component analysis and fisher linear discriminant analysis are successful face recognition algorithms. Recognition rate, time complexity can be improved by combining the two algorithms with the very powerful tool discrete wavelet transform. Experiments on the ORL face database show that the proposed method outperforms PCA, LDA, DWT+LDA algorithms in terms of recognition rate and classification speed. The proposed method is very powerful and useful in solving face recognition problems.
Keywords: 2D-DWT, PCA, 2D-PCA, LDA, face recognition.
Download Full-Text
ABOUT THE AUTHORS
Mostafa Mohamed Mohie El-Din
Department of Mathematics, Faculty of Science, Al-Azhar University, Cairo. Egypt
Mohamed Yousri El Nahas
Department of Systems Engineering & Computers, Faculty of Engineering, Al-Azhar University, Cairo. Egypt
Hassan Ahmed Hassan Elshenbary
Department of Mathematics, Faculty of Science, Al-Azhar University, Cairo. Egypt
Mostafa Mohamed Mohie El-Din
Department of Mathematics, Faculty of Science, Al-Azhar University, Cairo. Egypt
Mohamed Yousri El Nahas
Department of Systems Engineering & Computers, Faculty of Engineering, Al-Azhar University, Cairo. Egypt
Hassan Ahmed Hassan Elshenbary
Department of Mathematics, Faculty of Science, Al-Azhar University, Cairo. Egypt