A Comparative Study of DWT-HOG Based Face Recognition with Other Similar Techniques

The Discrete Wavelet Transform (DWT) is one of the efficient tools used in many image and pattern recognition applications. In this research, the DWT technique is combined with the Histogram of Oriented Gradients (HOG) to build a face recognition system. The proposal is to design and test the performance of the DWT-HOG based face recognition system and then compare the recognition rate results with other similar published techniques. To extract some useful sets of features from the addressed face image, the DTW technique is used first to extract the frequency content and provide the sub-bands for each of the selected face images. The HOG technique is then applied to each sub-band to extract the useful features. The performance of this new suggested system is measured by using the standard public face databases, ORL, AR and CMU-PIE. The experimental results proved that the proposed method is effective and relatively more accurate than other known published similar systems.