Abstract: In this study the method is presented for face image recognition identification and labeling. The method is based on using the combination of the Discrete Multi-Wavelet Transform (DMWT) and the Inverse Discrete Wavelet Transform (IDWT) followed by a Neural Network (NN). In this method the resulting coefficients were computed by the proposed multi-wavelets transform for single-level decomposition. It can be readily observed that the lowpass block (upper left corner) actually contains one lowpass subband and three bandpass subbands. The LL subbands resemble a smaller version of the original image. In this study the Inverse Discrete Wavelet Transform (IDWT) of LL coefficients will be obtained. The resultant of feature extraction is obtained by the Inverse Discrete Wavelet Transform, (IDWT) and applied to the first column to the Neural Network (NN) for recognition identification of the face image. This method gave an excellent result: 99% for a database of 50 different face images were excellent which indicates that the suggested algorithm in an excellent tool to process the database of standard pose of face image. A detection rate of 90% was achieved for facial movement such as a smile or other moves of face identification. The algorithm is implemented using MATLAB programming languages version 7.
Mikhled Alfaouri , Hilal M. Al-Bayatti and Nada N. Al-Ramahi , 2007. Novel Techniques for Face Recognition Identification and Labeling . International Journal of Soft Computing, 2: 216-224.