嵌入容量、图像视觉质量及不可检测性是衡量数据隐藏算法性能的重要指标。针对Ni等人提出的算法嵌入容量有限、安全性弱等问题,将矩阵分割技术及奇偶交错嵌入方式应用到数据隐藏领域中,提出一种简单、有效的可逆数据隐藏算法。该算法首先将图像矩阵分割为四部分,构建每部分的灰阶直方图,然后利用峰值集合中的像素来嵌入信息,由于增加了峰值集合中像素的数目,从而提高了嵌入容量。信息嵌入时,采用奇偶交错嵌入的方式,使得信息分布更不规律,从而增加了不可检测性。实验结果表明,相比Ni的算法,该算法在保证视觉质量的基础上,嵌入容量和不可检测性均有较大的提高。
Embedding capacity, visual quality of image and undetectability are the important parameters in evaluating a data hiding algorithm. To avoid the drawbacks of Ni algorithm, namely the low embedding capacity and low security level, in this paper we propose a simple and useful reversible data hiding algorithm by applying the matrix segmentation technique and odd-even interleaving embedding mode to data hiding. This algorithm first segments the image matrix into four parts and constructs the gray level histogram for each part, and then uses the pixels in peak point set to embed information, in this way, the number of pixels in peak point set is increased and the payload capacityis thus raised. In information embedding process, the algorithm adopts odd-even interleaving embedding mode, makes the distribution of the information embedded more irregular and thus enhances the undetectability. Experimental results show, compared with Ni algorithm, this algorithm significantly improves the embedding capacity and undetectability while assuring the visual quality of stego-image.