一种新算法在图像信息隐藏中的应用

摘要

A new algorithm based on the image DCT domain is proposed by using the DCT coefficients and HVS characteristic. Regarding the secret image scrambled by magic as bit streams and categorizing the sub blocks selected by random interval method, different embedding strategies are adapted to embed secret information. As for smooth sub blocks, secret bits are embedded by adjusting the relationship of intermediate frequency coefficients those have the same quantization coefficients. As for others, how many bits to embed are decided by the logarithm of the intermediate frequency coefficients. The secret bits are embedded in the DCT coefficients that being quantized and inverse quantized. Experiment results show that a large capacity of hiding information and a goodish robustness are achieved under the condition of imperceptibility. Moreover, the secret information can be extracted from the stego-image without referencing the original cover image.%在研究人类视觉特性和DCT变换的基础上,提出了一种基于图像DCT域的分块自适应隐藏算法.算法将经过幻方置乱的秘密图像转换为一维二进制流,并根据人类视觉特性将随机间隔法选择的载体图像子块分成三类,再分别采用不同的嵌入策略把秘密信息嵌入到子块中.对平滑块是通过调整两个量化系数相同的中频DCT系数的相对大小,来完成秘密信息的隐藏；对纹理块和边缘块是先通过量化系数的对数值来确定要嵌入的二进制流的位数,然后在经过了量化、反量化后的DCT系数上嵌入秘密信息.实验结果表明算法在保证良好的不可感知性的前提下,具有较大的信息隐藏容量及一定的鲁棒性,而且秘密信息的提取不需要原始图像的参与,实现了盲隐藏.