The Study And Realization Of Face Recognition Based On Android

AI has made significant progress in recent years, much attention has been paid to the biological recognition, face recognition as easy to operate, easy to accept interactive, safe and effective, without having to remember complicated account passwords and other advantages, has became one of the hottest direction in the pattern recognition research in recent years.At the same time, with the fast development of mobile Internet in recent years, a lot of daily business is moving online and bring more and more online authentication demand,due to the lack of the traditional username and password security, memorability and other aspects, especially with the increase in the number of user accounts, these drawback is becoming more and more obvious, therefore,the study of face recognition has from past PC gradually to the mobile terminal.Google’s Android operating system is released in 2008, a mobile device operating system, research on Android operating system is very convenient due to the openness of the it, catched the attention of domestic and foreign researchers.This paper discusses the mainstream algorithm of face recognition, described several algorithms in detail.improved kernel function of support vector machine and conduct experiments under the Android platform, main work and innovation points in this paper:(1)Analyzes of the demand in the current social recognition, analyzes the current market and future trend of the Android platform, analyzes development history of related algorithms of the process of face recognition include human face detection, feature extraction, classification and recognition, describes the current Android platform face recognition API services provided by the well-known internet company.(2)Analyzes the necessity of face image pre-processing, and discusses the influence of image normalization, histogram equalization, image smoothing process for the recognition accuracy and efficiency in detail and use OpenCV API to do experiment in Android platform.(3)Introduces the concepts and of face detection and feature extraction effects in the process of recognition, introduces the processes of Adaboost face detection algorithm in detail, including Haar features, weak classifier,strong classifier and cascade classifier; Analyzes the process ofPCA algorithm indetail; use OpenCV API to do experiments of face detection and feature extraction.(4)Introduces several common classifiers includes nearest neighbor classifier, bayes classifier and neural network classifier and support vector machine (SVM), lead to kernel function from the linear Linear Separable to the linear inseparable problem, this paper proposes a combined-kernel function that mixed sigmoid kernel function and radial basis function support vector machine classifier and experiment under Matlab, the experimental results show that the improved algorithm is better than the original algorithm. Analyzes the support vector machine classifier in face recognition under the multiple classification problems and completes the API of the demand.(5)Introduces Android application development core technology, OpenCV library, NDK, analyzes the principles and methods that using JNI calls OpenCV libraries in Android platform; analyzes the function of the system, system architecture, and system implementation.