In this paper, we propose a new reversible data hiding method in encrypted images. Due to spatial correlation, there is a large probability that the adjacent pixels of the image have small differences, which is especially obvious on the high four most significant bits (high nibbles) of the pixels. If the high nibble of each pixel is regarded as a 4-bit value, the differences between the high nibbles of the adjacent pixels are mostly concentrated in a small range. Based on this fact, Huffman coding was used to encode all the differences between the high nibbles of the adjacent pixels in order to compress the four most significant bit (MSB) planes efficiently and create a large-capacity room. After creating room, a stream cipher is used to encrypt the image, and the room is reserved in the encrypted image for data hiding without losing information. The experimental results showed that the proposed method can achieve a larger embedding rate and better visual quality of the marked decrypted image than other related methods.

本文提出了一种新的加密图像可逆数据隐藏方法。由于空间相关性，图像的相邻像素很可能具有小的差异，这在像素的高四个最高有效位（高半字节）上尤为明显。如果将每个像素的高半字节视为4位值，则相邻像素的高半字节之间的差异大部分集中在较小的范围内。基于此事实，霍夫曼编码用于对相邻像素的高半字节之间的所有差异进行编码，以便有效地压缩四个最高有效位（MSB）平面并创建一个大容量的空间。创建空间后，将使用流密码对图像进行加密，并且将空间保留在加密的图像中以隐藏数据而不会丢失信息。