On Mitigation of False Positive Problem in Singular Value Decomposition-Based Digital Image Watermarking

Abstract

Due to the effortless sharing of multimedia data (mainly images) over the communication channels and social media platforms, copyright protection of such data has become the topic of research interest. This paper presents a robust and false positive free watermarking scheme for digital image copyright protection with singular value decomposition (SVD) in the Stockwell transform domain. Unfortunately, most of the earlier SVD-based watermarking algorithms suffer from false positive or watermark ambiguity problem. In this study, we showed an approach to mitigate the false positive problem. A human visual system (HVS)-based method has been incorporated via discrete Cosine Stockwell transform (DCST). After projecting the input image in the DCST domain, SVD has been performed on the DCST coefficients. Finally, a total insertion-based approach has been presented to embed the watermark logo into the singular matrix of the DCST coefficients. Furthermore, the proposed approach’s security has been augmented by encrypting it with a novel cryptic algorithm called the chaotic logistics map. A detailed false positive analysis proves that the proposed method is free from all kinds of watermark ambiguity problems. Qualitative and quantitative comparisons with other state-of-the-art methods delineate the primacy of the proposed algorithm under most types of malicious signal processing and geometric impairments.