基于复合预测误差差值和互补嵌入的可逆数据隐藏

提出一种基于分块复合预测的误差差值和互补嵌入的彩色图像可逆数据隐藏算法。利用色彩分量之间的相关性和预测误差之间的关系减小差值,以增加差值直方图的峰值。采用双重嵌入方法提高嵌入容量,两次嵌入过程中像素值沿相反的方向扩展,部分像素值扩展量相互抵消,在图像分块的基础上选择最佳的预测方式组合,以增加扩展量抵消的机会,从而减小二次嵌入时图像质量的下降幅度。实验结果表明,本文算法在保证图像质量的同时大幅提高嵌入容量,算法的整体性能比其它同类算法更高。

Reversible data hiding based on the error difference of block compound prediction and the complementary embedding

This paper proposes reversible data hiding algorithm for color images based on block compound prediction and complementary embedding.The algorithm uses correlation of color components and the relationship between prodection errors to decrease difference,and the peak of difference histogram is increased.The embedding scheme makes use of the double embedding method to increase the embedding capacity,and uses the inverse direction to expand the pixel value,choosing the best combination of prediction method to increase the chance of expansion and cancel out each other.Experimental results show that the embedding capacity and the quality of stego-image are significantly improved compared with other algorithms.