Face Recognition Based on Wavelet Transform and Regional Directional Weighted Local Binary Pattern

With the development of information technology, face recognition technology has been continuously developed. This technology has attracted the attention of many researchers, including institutions and production enterprises. Face recognition technology has become a relatively independent application technology area in various social services. This paper presents a face recognition algorithm based on wavelet transform and regional directional weighted local binary pattern. First of all, this algorithm puts forward a new basis for a face recognition, namely the level of detailed components of face images containing valid facial texture details, and the recognition rate is better than that of the vertical component information and diagonal component information. This is called Horizontal Component Prior Principle(HCPP). According to HCPP, the original image is decomposed with wavelet transformation. The algorithm extracts the scale and level of detailed components. To improve the original LBP operator, it presents the regional directional weighted local binary pattern (RDW-LBP). Using the RDW-LBP, it can calculate the histogram of scale components and detailed components decomposed by wavelet. The histogram feature vector of face image can be got with the different weighted sub-regions. The feature vector can be matched with Chi-Square distance. This approach further enhances the ability to extract face direction information effectively.