IP语音隐写及隐写分析研究

随着IP语音(voice over IP, VoIP) 的快速发展，基于VoIP的隐写及其隐写分析逐渐成为信息安全领域的研究热点。总结和分析了近年来的IP语音隐写和隐写分析相关研究，并探讨该领域未来的研究方向。IP语音隐写根据其隐藏信息位置不同，可分为基于协议的隐写及分析和基于语音载荷的隐写及分析。前者是利用IP语音中的各类协议为载体，将信息嵌入到协议头部中的某些字段或通过调制IP语音包的时序来实现；后者则是通过修改语音载荷实现信息隐藏，包括基于最低有效位(least significant bit, LSB)的隐写，基于量化索引调制(quantized index modulation, QIM)的隐写和基于语音变化域回声等隐写。从协议和语音载荷两方面对IP语音分析技术进行了分类介绍。前者可分为协议头分析、时间序列分析、编码调制分析3类。后者包括基于特点语音参数（如线性预测参数、固定码本参数、自适应码本参数和所有参数）的隐写分析和基于语音信号处理的隐写分析方法。

Survey for voice-over-IP steganography and steganalysis

With the rapid development of internet protocol voice, voice over internet protocol (VoIP)-based steganography and steganalysis have grown as a hot research topic in the field of information security in recent years. This paper summarizes and analyzes the recent researches on IP voice steganography and steganalysis, and discusses the future research directions in this field. According to the location of the hidden information, IP voice steganography can be divided into protocol-based steganography and steganalysis, and voice payload-based steganography and steganalysis. The former is realized by embedding secret information into free or redundant areas of the protocol header or modulating the timing of IP voice packets with various protocols as carriers. The latter can hide secret information by modifying the voice payload, which can be divided into two categories, including steganography based on least significant bit (LSB) and steganography based on quantized index modulation (QIM). Correspondingly, the steganalysis of VoIP is classified into two aspects: the protocol-based steganography can be divided into three categories: protocol head analysis, time series analysis, and coded modulation analysis. The payload-based steganography includes steganalysis based on characteristic speech parameters (such as linear prediction parameters, fixed codebook parameters, adaptive codebook parameters, and other parameters) and steganalysis based on speech signal processing.