Research On 3D Face Recognition Based On Local Binary Pattern

Face recognition has a wide range of applications in many fields as a biology recognition technology,such as national information security,public safety,etc.Although the traditional 2D face recognition method can meet the requirements in some cases,the accuracy rate of recognition will be affected by the changes of illumination and pose.At present,3D face recognition technology has gradually been well known and has attracted extensive attention.,face recognition under complex circumstances has gradually become the focus of research in the field of face recognition.The key of face recognition is feature extraction,which directly affects the accuracy rate of face recognition.Local Binary Pattern is a very effective feature extraction method,which is widely used in face recognition technology.This paper mainly researches the following points based on LBP:Firstly,this paper deeply studies the problems of the traditional LBP algorithm and the improved algorithm of 3D face recognition.On the basis of this,local binary entropy pattern algorithm is proposed.Images with low resolution of RGB and depth(RGB-D)information are obtained by depth camera,and then the depth image is processed by entropy,which effectively enhance the resolution of depth images.Then the local and global mixed LBP algorithm is adopted for RGB-D image.The feature histogram of each region is obtained,which is matched with the face information in database finally,and face recognition is realized.Second,in order to reduce the impact of noise on face recognition further,this paper proposes a face recognition algorithm based on the frequency domain,which is based on the LBEP.Because there are different details in high and low frequency information of the image,this algorithm uses different filtering methods for the high and low frequency information,maximizing the retention of facial features while removing noise.Finally,combining the LBEP to match the feature information of face data.Thirdly,a face recognition system is designed and implemented on the basis ofimproved LBP algorithm.Each module of the system is analyzed and designed in detail by using the mixed programming method C++ and MATLAB.The system is mainly divided into five modules: data acquisition,face detection,feature point location,feature extraction and face recognition.The improved LBP algorithm proposed in this paper is used in feature extraction and face recognition modules.