A secure removable visible watermarking for BTC compressed images

A novel removable visible watermarking (RVW) algorithm by combining Block Truncation Coding (BTC) and chaotic map (RVWBCM) is presented in this paper. It embeds a visible watermark in the BTC codes of images, namely both the host image and the watermarked image are BTC compressed images. First, the original image is divided into watermarked region and non-watermarked region, and a predicted version of original image can be obtained by predicting pixel values in watermarked region. Second, adaptive embedding factors are computed according to the image features. Third, the watermark is adaptively embedded into two quantization levels of the BTC compressed image in visible manner. Meanwhile, to further prevent illegal watermark removal, original bi-level watermark is encrypted and then losslessly embedded in invisible manner by adjusting the relationship of two quantization levels. At the receiver’s end, only authorized users can exactly extract original bi-level watermark according the relationship of two quantization levels of BTC codes and succeed in remove the embedded visible watermark to reconstruct the original image. The experimental results show that this scheme can achieve a good balance between perceptual transparence and the watermark strength (watermark visibility) and can resist common image processing attacks. The proposed algorithm has low complexity and simplicity of implementation due to the use of BTC. It can be applicable to copyright notification and secure access control in mobile communication.