Title :
Steganalysis of MP3Stego
Author :
Qiao, Mengyu ; Sung, Andrew H. ; Liu, Qingzhong
Author_Institution :
Dept. of Comput. Sci., New Mexico Tech, Socorro, NM, USA
Abstract :
In this article, we propose a scheme for detecting hidden messages in compressed audio files produced by MP3Stego, as our literature search has found no previous work on successful steganalysis of MP3Stego. We extract moment statistical features on the second derivatives, as well as Markov transition features and neighboring joint density of the MDCT coefficients based on each specific frequency band on MPEG-1 audio layer 3. A support vector machine is applied to different feature sets for classification. Experimental results show that our approach is successful to discriminate MP3 covers and the steganograms generated by using MP3Stego.
Keywords :
Markov processes; audio coding; data compression; discrete cosine transforms; feature extraction; statistical analysis; steganography; support vector machines; MP3Stego; MPEG-1 audio layer 3; Markov transition features; audio files compression; modified discrete cosine transform coefficients; moment statistical feature extraction; neighboring joint density; second derivatives; steganalysis; steganograms; support vector machine; Computer science; Digital audio players; Digital images; Distortion measurement; Encoding; Feature extraction; Frequency; Steganography; Support vector machines; Transform coding;
Conference_Titel :
Neural Networks, 2009. IJCNN 2009. International Joint Conference on
Conference_Location :
Atlanta, GA
Print_ISBN :
978-1-4244-3548-7
Electronic_ISBN :
1098-7576
DOI :
10.1109/IJCNN.2009.5178971
