Reversible and robust image watermarking based on histogram shifting

In reversible watermarking, robustness of the watermark and the perceptual quality of the recovered host image has a major impact on the watermarking method. The proposed method provides improvement in the embedding capacity and the perceptual quality with a better robustness. Here, image pre-processing is performed by Gaussian filtering as a first step of the watermark embedding process. The peak points of the histogram are selected for embedding the secret data. In this method high frequency component modification is performed at the pixel position at which the watermark is embedded. A secret key is provided as an authentication process for the watermarked image, after adding the side information. At the extraction process, after the authentication and extraction of side information, Gaussian filter is applied. By using the side information the watermarked positions are identified, the secret data is extracted and the host image is recovered. The parameters such as embedding capacity, peak signal to noise ratio, Structural SIMilarity Index, bit rate and bit error rate are used for evaluation. The experimental results proves that, the proposed method provide better robustness and perceptual quality when compared with the existing method.