基于奇异值的鲁棒图像隐写算法

摘要

A image steganography algorithm based on singular value decomposition (SVD) and matrix encoding was proposed , which improved the robustness and imperceptibility of the data hiding method .After analyzing the stability of the singular value and using the EMD (exploiting modification direction) ,secret bits were embedded into the singular value obtained from the sin‐gular value decomposition of image blocks .The robustness of algorithm is insured by the singular value’s stability ,and the usage of EMD encoding method improves the embedding efficiency of the proposed algorithm .Experimental results show that the algo‐rithm has little embedding distortion ,and it also has better robustness compared with the original method .This method can be well applied in the image steganography in noise environment .%为提高数字图像隐写算法的鲁棒性和透明性，提出一种基于奇异值分解和矩阵编码的图像隐写算法。分析奇异值分解的稳定性，采用EMD （exploiting modification direction）矩阵编码方法，将秘密信息嵌入在将图像块进行奇异值分解所得的奇异值向量中。图像块奇异值的稳定性保证隐写算法的鲁棒性，采用EM D编码算法嵌入秘密信息比特使得算法具有较高的嵌入效率。实验结果表明，该算法嵌入秘密信息后图像失真小，与原始 EM D算法相比，该算法具有更好的鲁棒性。该算法能够很好的应用于噪声环境下的图像隐写技术。