Research On Face Recognition Using 2DPCA Based On Android Platform

With the further development of the mobile internet area, the popularity of intelligent terminal and its significantly enhanced computing performance,The actual application scenarios such as Mobile payment,phone unlocking,application unlocking on the mobile terminal are closely linked together with the face recognition technology and become a development direction.However,a key problem in face recognition techniques lies in the high dimension of face images.Researchers put forward the model of different algorithms based on subspace on the dimension reduction problem.C2DPCA algorithm is presented in this paper, and on the basis of the introduction of QR decomposition method,the C2DPCA method is extensed through the combination of it with traditional PCA,and Transplanting it to the android mobile terminal to implement the facial recognition application.The main works as follows:Firstly,the paper summarized the theoretical study and practice application development and the status of Face recognition technology domestic and overseas, Classify generally and analysis and compare the commonly used research algorithms in the filed of face recognition technologies,pointing out the advantages of PCA and 2DPCA variants relative to others on Transplanting to the android terminal.Then,it analyzed the application of face recognition using PCA,2DPCA and C2DPCA methods in the terms of cumulative eigenvalues,sample set Recognition rates and time consuming in Feature extraction relatively.Through experiments and expounds the differences and deficiencies.By introducing QR decomposition method,the C2DPCA algorithm is combined with the traditional PCA algorithm to Play their respective advantages of feature extraction for face images,and using Nearest neighbor classifier to do classification works.By the testing on different face databases, and got a better recognition rate and better computational efficiency.The last,designed and implemented an android application which principally aims on facial recognition as the main features.On the android mobile terminal.Based on the algorithm designed in this paper,On the platform using Java development language and C language and the Open source computer vision library which named opencv to implemente,and evaluate its function and performance indicators.Results show that in this paper, on the basis of the original algorithm performance, to further improve the recognition performance of face library and is conducive to the transplantation of mobile terminal.And on the android phones running results show that this method is applicable to the calculation process that applied on the resources limited of mobile devices.