Abstract
In this paper, we proposed a blind image watermarking scheme to enhance the effectiveness of hybrid Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) based schemes using the second level of SVD. The foremost contribution of this paper is to eliminate the constraints of our previous proposed scheme based on DWT and 2-level SVD. In the previous scheme, the size of the watermark was dependent to the size of cover image. For example, for a cover image with the size of 512*512, the watermark should be changed to 64*64 because of the limitation in capacity. In the proposed scheme, we enhanced the scheme to be independent from the sizes of cover and watermark images as well as improving the robustness and imperceptibility. For this purpose, we presented image blocking to find optimum block size for embedding the watermark and to correct the enslavement of capacity with the host image size. To achieve these goals, an additional preprocessing is performed on the watermark image before watermark embedding. The watermark image will be duplicated redundantly to the size of the cover image. Then both cover and watermark images will be divided into non overlap 16*16 blocks. Also, security is ensured in this scheme by proposing a two-level authentication system for watermark extraction to detect both false positive and false negative problems. The proposed scheme is tested and certified for medical and non-medical images exposing on 10 types of signal processing and geometric attacks. Experimental result shows, significant enhancement in both imperceptibility and robustness with the average of 46.9496 db and more than 99% respectively, while image blocking makes the capacity independent from the size of the host image. (C) 2019 Elsevier B.V. All rights reserved.