Imperceptible Image Watermarking based on Chinese Remainder Theorem over the Edges

Abstract

This paper introduced a watermarking method using the CRT and Canny Algorithm that able to improve the imperceptibility of watermarked image and preserving the robustness of watermark image as well. The classical CRT algorithm is spread the watermark bits evenly on the image area. It causes significant degradation when the embedding location lies on the least significant region or in the homogeny area. Otherwise, the proposed method embeds the watermark on the edges of the image which have significant difference value to maintain the imperceptibility. The Canny algorithm is used to indexing the embedding location based on the filtering output of host image. The watermark is then embedded into the host image using pair-wise coprime integers of 6 and 11 within the CRT modulo. The results show that the proposed method has significant improvement in the quality of watermarked image with the average value of 0.9995 compared to the CRT method which results in value of 0.9985. In compression and additive noise attacks the CRT has average values of 0.6618 and 0.9750, while the proposed method results in similar values of 0.6616 and 0.9752 respectively. These prove that the proposed method is able to preserve the robustness while improving the imperceptibility.