Reversible visible watermark embedded in encrypted domain

This work proposes a novel reversible visible watermarking scheme for encrypted images. In the scheme, the original plaintext image is encrypted by bit-wise exclusive-or operation. Although a data-hider does not know the original content, he may modify a part of encrypted data corresponding to the black pixels of a binary watermark image to insert the visible watermark. Meanwhile, some additional data used for content recovery are also embedded by using wet paper coding. After receiving the marked encrypted image, a user with the encryption key may perform direct decryption to obtain a visibly marked image, while a user with the encryption and data-hiding keys can perform data-extraction, image-decryption, and content-recovery to retrieve the original plaintext image.