Research On Face Recognition Algorithm Transplantation Based On Android Platform

In recent years,with the development of society and the progress of science and technology, the computer software and hardware performance rapidly increase. With the development of computer vision, pattern recognition and artificial intelligence theory and research framework, face recognition technology has become an important research direction of scientific research personnel concerned. At the same time, the progress of science and technology makes the mobile terminal has been greatly developed, compared to the initial stage, both have greatly improved the portability and computing and storage performance. Researching the traditional Face detection and recognition algorithm, combining with the characteristics of Android platform, select the appropriate algorithm improved after transplantation, which can promote the development of the application of face recognition under the Android platform. The application of face recognition on the Android platform, can be more convenient for the identification and authentication.This paper researches on the detection and recognition algorithms deeply, combining with the characteristics of Android platform, aiming at improving some methods, so that it can adapt to the Android platform, and then make the corresponding experiment and analysis. The work of this paper are as follows:(1) The research of face detection algorithm bases on cascade classifier under Android platform. This paper analyzes the weak classifiers, Adaboost strong classifier, cascade classifier, combining with the characteristics of Android platform, optimization of Haar features, designing face detection under the platform of Android, which laid the foundation for the later face recognition.(2) Researching on the face recognition algorithm based on Fisherface Android platform. Aiming at the characteristics of getting pictures simply but requiring high real-time performance based on Android, this paper analyzes the Fisherface algorithm based on the idea of Principal Component Analysis and Linear Discriminant Analysis. Compared with other face recognition researches on the platform of Android, and reducing the dimensionality reduction methods increasing the resolution, the Fisherface algorithm is transplanted to the Android platform for the research of face recognition, and little influencing the detection rate with reducing the amount of computation, and improve the efficiency of face recognition.(3) Studying the face recognition algorithm based on Local Ternary Plus Pattern feature under Android platform. By the light of the demand of Android platform combined with the characteristics of the single sample face recognition and image acquisition, this paper analyses the illumination of Local Binary Pattern features and Local Ternary Pattern features for better robustness, and according to the characteristics of Local Ternary Pattern. It puts forward a kind of method to describe image into uniform change more effectively named Local Ternary Plus Pattern feature, and studies face recognition algorithm under Android platform.(4) The research and implementation of Android bases on face detection and recognition system. According to the characteristics of Android platform, using cascade classifier for face detection, selects the Fisherface algorithm for face recognition in many samples to train, and under the conditions of single sample face recognition using LTPP features, and compares with some other face recognition system in the Android platforms.