Abstract: This paper analyses the Least Significant Bit(LSB) embedding scheme in wav audio and constructs the probability model of switch rule between wav audio’s sample values. Based on the probability model, it analyses the feasibility of using Chi-square detection, Sample Pairs Analysis (SPA) algorithm and Differential Image Histogram(DIH) algorithm to detect LSB steganography in wav audio. Experimental results show that the three algorithms can detect LSB steganography in 8-bit and 16-bit coded wav audio effectively. When the embedding ratio of this two coding manner is above 3% and 20% respectively, the correct rate of stego-audio’s judgments can reach 100%.

Key words: information hiding, steganalysis, Chi-square detection, Sample Pairs Analysis(SPA) algorithm, Differential Image Histogram(DIH) algorithm

摘要： 分析了wav音频中最低有效位(LSB)替换的嵌入机制，构造了wav音频采样点值在LSB替换中形成的值对转移概率模型，依据构造的概率模型分析了卡方检测、SPA算法和DIH算法在wav音频中检测LSB算法的可行性。实验表明，以上3种算法可对基于8比特和16比特wav音频的LSB替换进行有效检测，且对此两种编码方式，当嵌入率分别高于3%和20%时，对载密音频判决的正确率可达100%。

关键词: 信息隐藏, 隐写分析, 卡方检测, SPA算法, DIH算法

CLC Number: 

• TP391