Abstract：Existing instant voice communication steganalysis schemes are mainly based on supervised learning classifiers. These kinds of methods need large amounts of pre-processing and training and their accuracy can be easily destroyed by differences between the distribution of the training and testing data sets. This paper describes a semi-supervised hybrid detection model to improve detection which removes the manually annotated training data set, so this model is simpler and gives better detection scopes. This paper also describes a self-learning, multi-criteria fusion module which can automatically generate pseudo-labelled sets and combines the confidence and representative levels to judge the performance of instant voice communication steganalysis. There is no distribution mismatch between the testing data and the training data in this method. Tests with common low bit-rate speech coding carriers show that this method is more accurate than the un-supervised method and the supervised method in mismatched conditions. When the distributions of the training and testing data sets differ, this method is less affected than the supervised method. The tests also show that this method can be deployed on different kinds of speech codecs.
涂山山, 陶怀舟, 黄永峰. 基于半监督学习的即时语音通信隐藏检测[J]. 清华大学学报（自然科学版）, 2015, 55(11): 1246-1252.
TU Shanshan, TAO Huaizhou, HUANG Yongfeng. Detection of instant voice communication steganography using semi-supervised learning. Journal of Tsinghua University(Science and Technology), 2015, 55(11): 1246-1252.
[1] Mazurczyk W. VoIP steganography and its detection:A survey[J]. ACM Computing Surveys, 2013, 46(2):Article No. 20.
[2] Mazurczyk W, Karas M, Szczypiorski K. SkyDe:A Skype-based steganographic method[J]. International Journal of Computers Communications & Control, 2013, 8(3):432-443.
[3] Kopiczko P, Mazurczyk W, Szczypiorski K. Stegtorrent:a steganographic method for the p2p file sharing service[C]//Proceedings of 2013 Security and Privacy Workshops(SPW). San Francisco, CA, USA:IEEE Press, 2013:151-157.
[4] 两年前废掉了短信现又瞄准了语音通信[EB/OL].[2015-07-28]. http://ec.ctiforum.com/jishu/qiye/qiyetong- xinjishu/jishitongxin/jishudongtai/433729.html. Two years ago WeChat gave up SMS, now is aiming at voice communication[EB/OL].[2015-07-28]. http://ec.ctiforum.com/jishu/qiye/qiyetongxinjishu/jishitongxin/jishudongtai/433729.html.(in Chinese)
[5] 看极端组织ISIS如何展开网络"恐怖营销"[EB/OL].[2015-07-28]. http://www.techweb.com.cn/news/2014-09-01/2070887.shtml. See the extreme organization ISIS how to expand the network "terrorism marketing"[EB/OL].[2015-07-28]. http://www.techweb.com.cn/news/2014-09-01/2070887.shtml.(in Chinese)
[6] LI Songbin, TAO Huaizhou, HUANG Yongfeng. Detection of quantization index modulation steganography in G.723.1 bit stream based on quantization index sequence analysis[J]. Journal of Zhejiang University SCIENCE C, 2012, 13(8):624-634.
[7] TIAN Hui, LIU Jin, LI Songbin. Improving security of quantization-index-modulation steganography in low bit-rate speech streams[J]. Multimedia Systems, 2014, 20(2):143-154.
[8] HUANG Yongfeng, TANG Shanyu, BAO Chunlan, et al. Steganalysis of compressed speech to detect covert voice over Internet protocol channels[J]. IET Information Security, 2011, 5(1):26-32.
[9] BAO Chunlan, HUANG Yongfeng, ZHU Chunyi. Steganalysis of compressed speech[C]//Proceedings of Multiconference on Computational Engineering in Systems Applications. Beijing, China:IEEE Press, 2006:5-10.
[10] LIU Qingzhong, SUNG AH, QIAN Mengyu. Temporal derivative-based spectrum and mel-cepstrum audio steganalysis[J]. IEEE Transactions on Information Forensics and Security, 2009, 4(3):359-368.
[11] Koçal OH, YürüklüE, Avcibas I. Chaotic-type features for speech steganalysis[J]. IEEE Transactions on Information Forensics and Security, 2008, 3(4):651-661.
[12] Kraetzer C, Dittmann J. Pros and cons of Mel-cepstrum based audio steganalysis using SVM classification[J]. Information Hiding Lecture Notes on Computer Science, 2007, 4567:359-377.
[13] Janicki A, Mazurczyk W, Szczypiorski K. Steganalysis of transcoding steganography[J]. Ann. Telecommun. 2014, 69(7-8):449-460.
[14] HUANG Yongfeng, TANG Shanyu, ZHANG Yuan. Detection of covert voice-over Internet protocol communications using sliding window-based steganalysis[J]. IET Communications, 2011, 5(7):929-936.
[15] HUANG Yongfeng, LIU Chenghao, TANG Shanyu, et al. Streganography integration into a low-bit rate speech codec[J]. IEEE Transactions on Information Forensics and Security, 2012, 7(6):1865-1876.
2015, Vol. 55 
