摘要
传统的身份认证技术无法对云端的海量访问认证请求进行有效的管理。目前的多因素认证机制中,用户的验证码由云服务提供商支持的多因素认证设备免费生成,而基于专门的动态口令卡的多因素认证服务则需要收取一定的费用。为了实现免费身份认证,采用一种区分性模型方法实现声纹识别说话人身份的确认,并采用信道补偿算法进一步提高系统性能。做了与生成性模型声纹识别身份认证技术进行对比和分析的实验,得到新方法具有鲁棒的身份认证性能的结论。
Traditional certification technology can not effectively manage the cloud mass access certification request. In the muhi-factor certification mechanism, the user＇s verification code which multi-factor certification device generates at no charge is provided by cloud service providers. And multi-factor certification service based on the specialized dynamic password card needs to charge fees. This paper presents one free certification method with voiceprint recognition. The speaker identification task can be achieved by the discriminative model and channel compensation algorithm to further improve the system performance. Through the comparison and analysis of experiments of the generative model voiceprint identification certification technology, the new method has robust certification performance.
出处
《现代电子技术》
2013年第3期98-101,共4页
Modern Electronics Technique
关键词
气象业务
云服务
多因素认证机制
声纹识别
区分性模型
meteorological service
cloud service
multi-factor certification mechanism
voiceprint identification
discriminative model
