Face recognition based on DWT/DCT and SVM

Discrete wavelet transform has both good qualities in time domain and frequency domain which is an ideal tool in analyzing unsteady signals. Discrete cosine transform is one of the approaches used in image compressing which is also used to extract features. This paper proposes a combined feature extraction method which is based on DWT and DCT for face recognition. First the original face image is decomposed by 2-dimentional DWT, then the 2-dimentional DCT is applied to the low frequency approximation image obtained from previous step. In the end, using the DCT coefficient, a SVM classifier is built and face image can be recognized. The experiment carried on ORL-DATABASE shows that the above-mentioned feature extraction method can gain higher recognition rate than the traditional PCA algorithm.