基于语音特征的身份认证机制研究

发布时间：2018-10-30 06:47

【摘要】：身份认证作为网络安全的第一道屏障,已是目前网络安全中最为热门研究课题之一。传统的基于口令的身份认证,一旦口令被盗用,服务器很难判断用户身份的真伪,所以需要对传统的身份认证进行改进,来增加其安全性。由于语音特征具有唯一性、安全性、稳定性、可靠性等优势,使其在身份认证中得到广泛应用。本文针对语音特征的这些优点,在传统的身份认证的基础上提出基于语音特征的身份认证机制,将语音特征作为传统身份认证中的语音验证码,进而增加了身份认证的安全性。本文通过对语音信号进行预加重、分帧加窗、端点检测、特征的提取等操作完成对语音信号的分析,提取了语音特征LPC、LPCC、MFCC。本文将语音特征与SHA-1算法相结合完成对用户的消息鉴别,解决了身份认证的完整性和不可否认性两大难题。并通过DTW语音识别算法和结合语音验证码的挑战-应答机制解决了身份认证中的可信性和访问控制两个难题。最后,本文在VC6.0的环境下,完成了对基于语音特征的SHA-]消息鉴别算法的测试,将“0-9”十个数字语音作为输入,比较输出的散列值是完全的不同的,得出结论将语音特征作为消息鉴别的一个标识是非常安全可靠的。并在MATLAB的平台下完成了对DTW语音识别算法的测试,分别对不同环境下的“0-9”十个数字语音的识别率进行测试,测试结果表明DTW算法的对“0-9”十个数字的识别率是非常高的。证明了语音特征作为身份认证的一个验证标识是切实可行的,最后本文结合语音特征和挑战-应答机制提出了将”0-9”十个数字中任意六个数字的随机组合作为身份认证的语音验证码。
[Abstract]:As the first barrier of network security, identity authentication is one of the most popular research topics in network security. The traditional identity authentication based on password, once the password is stolen, it is difficult for the server to judge the authenticity of the user identity, so it is necessary to improve the traditional identity authentication to increase its security. Speech features are widely used in identity authentication because of their uniqueness, security, stability and reliability. Aiming at these advantages of speech features, this paper proposes an identity authentication mechanism based on speech features on the basis of traditional identity authentication, which takes speech features as the voice verification codes in traditional identity authentication, and then increases the security of identity authentication. In this paper, the speech signal is analyzed by pre-weighting, framing and windowing, endpoint detection, feature extraction and so on. The speech feature LPC,LPCC,MFCC. is extracted. In this paper, the combination of speech features and SHA-1 algorithm is used to complete the message authentication of users, which solves the integrity and non-repudiation problems of identity authentication. The credibility and access control problems in identity authentication are solved by using DTW speech recognition algorithm and the challenge response mechanism of speech verification code. Finally, under the environment of VC6.0, this paper completes the test of the SHA-] message discriminant algorithm based on speech feature. Taking "0-9" ten digit speech as input, the hash value of output is completely different. It is concluded that it is very safe and reliable to use speech feature as an identification of message identification. The DTW speech recognition algorithm is tested on the platform of MATLAB, and the recognition rate of "0-9" digital speech in different environments is tested respectively. The test results show that the recognition rate of "0-9" 10 digits by DTW algorithm is very high. It is proved that it is feasible to use speech features as an authentication mark for identity authentication. Finally, combined with speech features and challenge-response mechanism, this paper proposes a random combination of any six of the "0-9" ten digits as a speech verification code for identity authentication.
【学位授予单位】：黑龙江大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN912.3

【相似文献】