Research And Implementation Of The Voiceprint Recognition System Based On Android Platform

With the social progress and economic development, the popularization and application of computer technology, has penetrated into all aspects of national life. In such a highly interactive information society, determining a person’s identity is a very important issue, especially in the fast-paced information age, the need to identify and interactive applications is increasingly widespread, the required to verify the identity of the human is more and more. Therefore, authentication of digital, hidden, easy and convenient are becoming increasingly important. Using the human voice in daily life to communication is the most convenient, most common and most effective form of communication. Language is a human-specific features, the human voice is the most commonly used tools. Allow computers to understand human language, to achieve human-computer interaction is people’s dreams since the birth of the computer.Voiceprint Recognition, also known as Speaker Recognition, is a technology to distinguish the speaker’s identity by voice.It belongs to the scope of the biometric technology as fingerprint recognition which people are familiar with, and it is easy to use, will not be lost.In the highly developed information technology society, automation and intelligence has become increasingly common, to identify and determine who a man is very important. However, the traditional identification and determination has lagged behind, and far from being able to meet people’s needs. The technology of identification based on the biological characteristics provides us with a convenient, efficient and safe method.The key technology of voiceprint recognition and realization is the speech signal feature extraction and pattern matching. The paper discusses in detail the cepstrum feature extraction method based on acoustic characteristics, the VQ voiceprint identification based on model matching, and design a VQ-based text-independent voiceprint recognition system in the Android platform.The main content is as follows:(1) This paper studies the process of the speech signal, especially analysis the feature extraction of speech signal;(2) This paper studies the basic principle of voiceprint identification, voiceprint identification process, including pretreatment, endpoint detection, feature extraction, modeling, pattern matching algorithm;(3) Combined with practical application, we design a vector quantization algorithm based on the mobile platform, the algorithm can effectively solve the voiceprint recognition the problem of refusing to identify and we simulate the algorithm;(4) Using server and client mode, we design and implement the voiceprint recognition system based on the Android platform, users can use it online or offline, and the system identification speed and accuracy is high.In the next few decades, advanced communications will rapidly change the way people live and work, anywhere and anytime using a range of equipment for real-time communication and data processing will gradually become a reality, mobile computing platform based on voiceprint recognition is one of key technologies. Voiceprint recognition in mobile commerce, distance learning, video surveillance, automatic authentication, human-computer interaction, financial security, and judicial investigators and other fields, will have important applications.