Research On Detecting Of Audio Steganography Techniques Based On Statistics

As the rapid development of informationization in government and army, the problem ofinformation security is becoming increasingly prominent. The Steganography provides a newmeans for information security and also provides criminals a confidential way to send illegalmessages. Steganalysis is just a confrontation techniques against steganography, can effectivelymonitor steganography being illegal malicious applied and can reveal the current deficiencies insteganography, to promote the development and improvement of steganography. The Detectiontechnique of steganography is the first step of Steganalysis, is the premise and foundation ofextraction and destruction of steganography. Audio steganography is one of the most likely to actas carrier, therefore the research of audio steganography detection technique is of greatsignificance. In this paper, the main work including:1. In-depth analysis of the audio steganography and steganalysis technique is carried on. Itis pointed out that the stego-detection that aims at audio frequency domain QIM steganographyand space-time domain LSB matching steganography is the weakness in audio steg-detection.2. A new stego-detection method aims at audio frequency domain QIM steganography isresearched. Based on the the mechanism of the QIM stego-operation, we analyse the change onthe statistical characteristics of audio covers. Through the design of the statistical characteristicsquantity, the characteristics of the stego-audio are reflected. On this base, a specific detectionalgorithm is proposed. Algorithm analysis and experimentation result show the low complexity,the high detection rate, and the good applicability of this algorithm.3. A new stego-detection method aim at space-time domain audio LSB matchingsteganography algorithm is researched. First, based on the analysis of the mechanism andexperimental, the characteristics of stego-audio are discovered, and the theoretical analysis andreasoning shows the characteristics of stego-audio are universal. Then a statistical featurequantity that is sensitive to steg-operations and can also reflect the characteristics is designed.Based on this a steg-detection algorithm is proposed. Algorithm analysis and experimentationresult show that the algorithm is a peed, high detection rate and low complexity.4. A stego-detection system of audio is designed and implemented.Finally, sum up the work in this thesis, and the development prospects and researchdirections of the audio steganography detection technology are suggested.