Abstract: A new reversible data hiding algorithm is proposed for low embedding rate and complexity process in traditional algorithm. To increase the information hiding rate, the highest bins at both sides of the median point in image histogram are chosen to be split for data embedding, respectively. Multiple pairs of histogram bins are chosen for data embedding and the pair of the median point is chosen at last. The median point can be located directly after embedding, so that reversible data embedding and image recovered can be achieved, in addition, there is no need for the replacement of the least significant bit which requires a simple computation. The proposed algorithm is tested on two image sets. The obtained experimental results are compared with the existing algorithm for data embedding and  three MATLAB functions for image contrast enhancement, which can evaluate the capacity of embedding and the quality of image contrast enhancement. The numerical results show that the proposed algorithm outperforms the existing algorithm in hiding rate and image quality preservation, and besides, the capacity of data embedding is greatly improved.

Key words: reversible data hiding, image contrast enhancement, embedding rate, median point, histogram peak

摘要： 针对已有算法嵌入率低、算法过程复杂等问题，提出一种新的可逆数据隐藏算法。算法为了提高信息隐藏率，首先选取直方图中值点两侧的峰值进行数据嵌入。在多次嵌入之后，算法选择中值点位置嵌入所有峰值信息，由于中值点在嵌入后可以直接定位，因此可以实现可逆嵌入与恢复，从而不需要已有算法最低有效位的替换过程，减少了算法过程的复杂性。算法在两个图像集上进行性能测试，先与已有算法进行数据嵌入量比较，接着与MATLAB中3个对比度增强函数进行比较，用于评估算法的嵌入容量以及对比度增强后的图像质量。实验结果表明，所提算法在实现可逆数据隐藏和对比度增强的同时，嵌入容量有较大的提高。

关键词: 可逆数据隐藏, 图像对比度增强, 嵌入率, 中值点, 峰值