Content-adaptive reversible visible watermarking in encrypted images

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