Content-adaptive reversible visible watermarking in encrypted images

摘要

The reversible visible watermark which serves as the perceptual ownership identifier can be extracted to losslessly recover the original cover media. This paper presents a novel content-adaptive reversible visible watermarking scheme in encrypted images. To achieve the tradeoff between watermark visibility and marked image quality, data embedding positions for accommodating the watermark are adaptively selected using the visual perceptual model before encryption. Due to weak spatial correlation in encrypted images, the data embedding room is vacated before encryption with a traditional reversible data hiding algorithm to contain pixel bits in data embedding positions. Therefore, it is convenient for the data hider to embed the visible watermark in encrypted images by substituting pixel bits in data embedding positions. If the receiver decrypts the marked encrypted image without extracting the embedded watermark, the visibly marked image can be obtained. In addition, if the receiver decrypts the marked encrypted image and extracts the embedded watermark, the original image can be perfectly recovered. Experimental results demonstrate the merits of the proposed scheme in terms of marked image quality, watermark visibility and watermark robustness. (C) 2019 Elsevier B.V. All rights reserved.