基于汉明距离统计的隐写分析算法

Steganalysis Method Based on Hamming Distance Statistics

针对LSB匹配隐写算法会改变自然图像LSB(最低有效位)序列相关性的特点,提出了"汉明距离统计"方法。在合理假设的情况下,分别对随机序列和自然图像LSB序列的汉明距离统计分布进行推导和实验,构造了能够度量二者相关性本质差异的统计量,并据此设计了适用于LSB匹配的隐写分析算法。实验结果表明算法具有较好的检测性能。

Aiming at the fact that the LSB-Matching steganographic algorithm changes the correlation of the natural image LSB（Least Significant Bit） sequence, this thesis proposed a statistics method called ＂Hamming Distance Statistics＂. With reasonable assumption, specific derivation and clear experiment were executed to the Hamming Distance statistics results of natural image LSB sequence and random sequence respectively, and we founded a statistic to measure the essential correlation difference between them. With the statistic we designed a steganalysis method applicable for detecting LSB -Matching steganographic algorithm. The experimental results showed satisfying detection performance.