Lossless data hiding in JPEG bitstream using alternative embedding

This paper proposes a lossless data hiding method for JPEG images using adaptive embedding. By constructing an optimal mapping between the used and unused Huffman codes in each category, we take full use of the combination of mapping to achieve a high embedding rate. In order to improve the payload, we further use a code reordering based embedding algorithm. Both algorithms are alternatively used during data hiding. After modifying the Huffman Table defined in JPEG header and substituting the codes in entropy-encoded segments, additional messages are embedded into the JPEG bitstream. The proposed method is lossless to the image, i.e., the decoded content of a marked JPEG bitstream is identical to the original JPEG image. Meanwhile, the file size can be well preserved after data hiding. Experimental results show that the proposed method has a better performance than state-of-the-art works.