摘要
为了在保证语音通信实时性和一定隐藏容量的同时降低隐写失真以达到较好的不可感知性,提出了一种以低速率语音为载体的双层隐写算法。通过限制语音子帧基音周期值的搜索集合,实现了第一层隐写;利用搜索集合内基音周期取值的任意性,实现了第二层隐写。双层嵌入过程中,以最小化修改幅度为原则决定基音周期的取值,降低了隐写失真。实验结果表明,算法在实时性方面性能优越,并可较好地抵抗隐写分析算法的检测,在嵌入率达到100 bps时仍保持了良好的语音听觉质量。
In order to reduce the steganography distortion to achieve high imperceptibility while keeping a good real-time per- formance and a suitable hiding capacity, this paper proposed a double-layer steganography algorithm for low bit-rate speech. By limiting the search range of the pitch periods in the subframes, it realized the first layer embedding. Utilizing the characte- ristic of pitch periods getting values in the search range optionally, it realized the second layer embedding. During the double- layer embedding process, it selected the values of pitch periods in terms of minimizing the modification amplitude, which re- duced the steganography distortion. Experimental results show that while achieving satisfactory performances in the aspects of real-time and anti-detection, the proposed algorithm can keep a good speech quality at an embedding rate of 100 bps.
出处
《计算机应用研究》
CSCD
北大核心
2015年第6期1774-1777,共4页
Application Research of Computers
基金
河南省科技攻关计划资助项目(122102210047)