Lossless data hiding for VQ indices based on neighboring correlation

In this paper, a novel lossless data hiding scheme is proposed to achieve the goals of hiding secret data into a VQ compressed image and lossless reconstruction of the original image. This scheme first uses neighboring and processed compression codes to generate sub-codebooks and then uses these sub-codebooks and encoding algorithm to perform encoding and data hiding simultaneously. More important, the original VQ compressed codes can be recovered completely after data extraction. To evaluate the effectiveness of this approach, various test images are used in the experiments. From the experimental results, it is shown that the proposed scheme can largely decrease the compression rate and increase the hiding capacity for achieving data hiding. That is, the performance of the proposed scheme is superior to other previous schemes in compression ratio and embedding capacity.