Abstract: According to the existing algorithm’s low embedding capacity and unstable for medical image, a novel reversible data hiding method in encrypted medical images based on Slash grouping and gradient operator is proposed. After encrypting the medical image by bitwise exclusive-or operator, the encrypted image is divided into blocks and each block contains four specific groups. Then each block embeds two bits by flipping three LSBs of corresponding group. With an encrypted medical image containing additional data, it is firstly decrypted using the encryption key. Then the image recovery and data extraction can be achieved according to the data-hiding key and the gradient operator. The experimental results show that the original medical image can be perfectly recovered and the embedded data can be successfully extracted. When the block size is the same, the proposed method offers better embedding capacity, higher PSNR and lower extracted-bit error rate.

Key words: encrypted medical images, reversible data hiding, image recovery

摘要： 针对现有算法嵌入容量低以及应用于医学图像时效果不稳定的问题，提出一种基于斜线分组和梯度算子的密文医学图像的可逆信息隐藏算法。算法利用按位异或操作对医学图像加密，接着对密文图像分块，每块再斜线分为4组，根据待嵌入2比特信息翻转对应像素组的最低3位有效位（3LSBs）。接收端先解密图像，然后根据信息隐藏密钥和梯度算子恢复原始图像并提取信息。实验结果显示，算法能无损地恢复原始医学图像，准确地提取嵌入信息，在分块大小相同的情况下，有更好的嵌入容量，更高的PSNR值和更低的信息提取错误率。

关键词: 密文医学图像, 可逆信息隐藏, 图像恢复