Study On Voice-over-IP Steganography Using Adjustable Matrix Encoding

Voice over IP(Vo IP) as a typical example of streaming media, it can provide favorable conditions for information hiding, which have caused widespread concern of information security researchers. However, the real-time of Vo IP, limited carriers of each voice frame and the non-equilibrium impact of voice carrier, which bring some great challenges to digital steganography. So, this paper made a further research on the basis of traditional matrix encoding, present an adjustable matrix encoding approach. Based on this theoretical method, explore many new Vo IP steganographies to cope with the challenges. The main contributions of our study are as follows.(1) Traditional matrix encoding is limited by the length of carrier, which leads to low applicability. Considering that, a new adjustable matrix encoding(AME) is proposed, which idea is that construct a basic matrix by two hamming check matrix, then randomly expand its column to obtain a new guide matrix which can adapt to the needs of different carrier of length and improve the steganogrphic performance. Theories and experiment demonstrate that adjustable matrix encoding has a better applicability than matrix encoding, and has high steganogrphic performance at the same time.(2) For reducing the computational complexity of Wet paper code(WPC) and meeting the real-time requirement of Vo IP steganography, a steganographic method based on improved WPC is proposed, which basic idea is that based on AME to improve the WPC, reversely constructing the model of WPC, which can reduce computational complexities to O(n) and obtain the optimal solution. Experimental results demonstrate that the proposed method can satisfy the real-time of Vo IP and has high steganographic performance.(3) For obtaining the high steganographic efficiency in the carrier of arbitrary length, a steganographic method based on optimal matrix encoding is proposed; and divide-and-rule strategy is adopted, which often divide secret message and cover into many small parts, use AME to satisfy the length of each small part and obtain the higher steganographic efficiency; then, the optimal embedding problem is converted to a typical combinational optimization problem, which can obtain the optimal resolution. Theory and experiment show that the proposed method can obtain the optimal embedding efficiency in any case.(4) The carrier of voice has the non-equilibrium characteristics, which are ignored will cause additional distortion for voice carrier. To resolve this problem, a non-equilibrium steganographic method is proposed, which idea is that AME can changes the carrier with different probability can change the carrier, which has a small impact on voice, with a large probability, and change the carrier, which has a large impact on voice, with a small probability, further reduce the distortion of carrier. Experimental results show that when bit change rate is same, non-equilibrium steganography can cause a less distortion to carrier than equilibrium steganography.