An improved VLC-based lossless data hiding scheme for JPEG images

•We propose an improved VLC-based lossless data hiding scheme for JPEG images.•The main improvement is to optimize the establishing of mapping relationships, so as to further improve the embedding capacity.•We extend the proposed optimization algorithm to the situations in which slight file size increase is allowed and get higher embedding capacity.•Our scheme keeps exactly the same image content as the original one after embedding.

In this paper, a lossless data hiding scheme which directly embeds data into the bitstream of JPEG images is presented. For real cases, the JPEG code space is partly occupied, not all variable length codes (VLC) in the Huffman table are used during the JPEG image compression process. Thus, these unused VLCs can be made used of and data hiding can be performed by mapping one or more unused VLCs to one used VLC. Through analyzing the statistics of both used and unused VLCs, the proposed scheme can take full advantage of the unused VLCs by mapping Huffman codes according to a specific mapping strategy and reach higher capacity. The output stego image can keep exactly the same content as the original one and preserve the same file size, and if the file size is allowed to be enlarged, then our scheme can achieve a significant improvement of embedding capacity.