Font Size: a A A

Research On Android Platform Voiceprint Unlock System

Posted on:2017-02-22Degree:MasterType:Thesis
Country:ChinaCandidate:X X HaoFull Text:PDF
GTID:2348330482486484Subject:Control engineering
Abstract/Summary:PDF Full Text Request
Mobile Internet is gradually penetrate into people's life, work in various fields,and handheld devices playing an increasingly important role in people's lives. Phone call, send SMS, MP3, camera and other functions already can not meet the modern needs of today's mobile platforms for increasingly updated, more in line with consumer demand, such as online shopping, transfers, mobile banking and other functions, through people phone and the outside world to complete the exchange all kinds of information, and therefore mobile security and confidentiality of the phone itself more and more important.Combining the research hotspot authentication technology, mainstream operating system Android phone system, we put forward a biometric authentication technology voiceprint recognition technology used in mobile phones owner identity authentication, and directed to the voiceprint recognition technology Android system internet-related issues involved in the analysis and research.In this paper, the main innovation lies in:(1) the use of mobile phones anytime,anywhere, at any time may unlock the phone using a variety of functions, since the complex environment, various noises will affect whether or not able to accurately unlock the phone. For noise environment speech enhancement technology research,and in-depth study to unlock algorithm. In order to improve the robustness of the algorithm in the phone application environment noise, by experimentally determining LMS adaptive filter is a voiceprint noise reduction algorithm; and noisy speech issues for the traditional method of endpoint detection improvements.(2)when the phone is unlocked, the password normally not too long, cumbersome lengthy owner password is easy to forget the lack of medical users, taking into account the limitations of the Android mobile phone computing power, and considering the time spent voiceprint recognition algorithms to identify and balancerate, and focuses on the dynamic time-related structured text(DTW) voice for small libraries and vector quantization(VQ) classification methods, experimental data comparing the voiceprint identification result.(3) paper implements voiceprint lock based on the Android platform, requires not only a person's voiceprint information when you unlock the phone, you also need to provide the correct semantic information double protection, so, even if the phone is stolen, your phone Important Information not to divulge.
Keywords/Search Tags:Voiceprint unlock, Android system, Feature extraction, Endpoint detection, Speech enhancement
PDF Full Text Request
Related items