A high-capacity reversible data hiding scheme for encrypted images employing vector quantization prediction

In this paper, a high-capacity RDHEI scheme based on Vector Quantization prediction (VQP) and adaptive block selection is proposed. VQ compression is a simple lossy compression method. In our scheme, VQ compression is used to estimate the original pixels to vacate room before encryption. Since the difference between the VQ decompressed image and the original image is small when the length of codebook is sufficient, using the difference as a prediction error can obtain more space for data embedding. Moreover, adaptive block selection is used in our scheme. Each block in the image dynamically adopts different data embedding strategies according to different block types. In addition, the experimental results show that our scheme can not only obtain considerable embedding capacity and perfect image restoration effect, but also has a good performance in terms of security.