Study On Biological Information Based Android Identity Authentication

With the rapid development and popularity of the Internet, the security problem of the data information in mobile terminal device attracts more and more attention. According to the characteristics that users access to the information stored in mobile devices, such as contacts, text messages, photos, all kinds of account password in high frequency, on the premise of ensuring the safety, identity authentication and authorization is a more efficient and convenient way than data encryption and decryption. However, many disadvantages still exist in traditional digital, grap hics and smart card authentication method, such as easy to be forgotten, lost, stolen and inconvenient to be carried. Because biological information has many advantages for authentication, such as could not be forged, no need to remember, easy to carry and easy to verify, the method associated biological information with sensitive information can improve the security of sensitive information without compromising ease of use.Two authentication systems based on voice and facial bioinformatics features are studied and designed in this thesis. For the problem that voice and face mode authentication need handle large amounts of data and the calculation and storage capacities of the mobile terminal device are limited, a technology algorithm that suitable for mobile terminal platform is proposed in this thesis by studying the relevant technical methods and optimizing the parameters in algorithm. It improves the security level of the system by combining traditional content and voiceprint authentication in voice authentication module; Using the voiceprint characteristics update voice password training template dynamically after each successful authentication, which reduces the impact of the time- varying characteristics of voice, and improves the robustness and adaptability of the authentication system. Test result shows that 5 is best training times to get the voice password training template and the best overall effect in the thesis. To reduce the amount of computation, improve the data processing efficiency, and weaken the influence of the background for image detection, the face image acquisition window is designed in face authentication modules. The root key design is added to the system to improve system security at the same time enhance the system use convenience.The goal of this thesis is to study and design a secure, convenient, efficient, and reliable authentication systems base on bioinformatics voice(voiceprint + content) and facial features. For this goal, the thesis does the following works:1. To complete the basic preparation of the theoretical, the principles of voice(voiceprint + content) and face authentication and the main implementation technology in each stage have been studied by looking through a lot of literatures. And the implement solution of the voice(voiceprint + content) and face authentication on the Android platform has been analyzed. 2. The proposed system focuses on the study on voice preprocessing, MFCC(Mel Frequency Cepstrum Coefficient) feature extraction and Gaussian mixture model(GMM) pattern matching and other technical methods in the system voice authentication module. 3. The proposed system focuses on the study on the face detection based on Haar classifier, face authentication based on Fisher linear discriminant threshold settings and other technical methods in the system face authentication modules. 4. The voice(voiceprint + content) and face security authentication system based on the Android platform is designed and implemented. The problems raised in the analysis are solved. The thesis testes and analyzes the main technical indicators to verify the effect of the realization, and compares with the same type of authentication system using different methods. The result shows that the proposed system has higher security and convenience.