Abstract: Based on the characteristics which reversible data hiding technology can accurately extract the embedded secret information and restore the original carrier. It can be potentially used in the medical and military applications. The paper proposes a reversible data hiding algorithm based on histogram shift. According to the characteristics of pixel distribution in medical images, the Method of Maximum Classes Square Error is used to separate the background and the foreground to get the foreground contour. And the pre-process in an existing algorithm is improved to adaptively select the optimal pre-process interval, in the information embedding procedure, the contrast of the region of interest can be gradually enhanced. Furthermore, the original image can be exactly recovered. Compared with the previous method, in this paper, the contrast enhancement effect of medical image ROI is more obvious and the visual quality is greatly improved.

Key words: reversible data hiding, medical image, histogram shift, background segmentation, region of interest, contrast enhancement

摘要： 基于可逆数据隐藏技术具有可以准确地提取嵌入的秘密消息并无损还原原始载体的特点，可逆数据隐藏技术被广泛使用在医学和军事等领域。提出一种新的基于直方图平移的可逆数据隐藏算法，针对医学图像的像素分布特点，利用最大类间方差法分离背景和前景区域，确定边缘像素。对已有算法的预处理过程进行改进，使其能够自适应地选择最优预处理区间，降低视觉失真，并在信息嵌入的同时使医学图像感兴趣区域的对比度增强效果得到逐步改善，最终原始医学图像可以无失真地恢复。与现在已知的方法相比，新算法可以使医学图像ROI的对比度增强效果得到更明显的改善，大幅度提高了图像的视觉质量。

关键词: 可逆数据隐藏, 医学图像, 直方图平移, 背景分割, 感兴趣区域, 对比度增强