A Novel Multipurpose Watermarking Scheme Capable of Protecting and Authenticating Images With Tamper Detection and Localisation Abilities

Technologies that fall under the umbrella of Industry 4.0 can be classified into one of its four significant components: cyber-physical systems, the internet of things (IoT), on-demand availability of computer system resources, and cognitive computing. The success of this industrial revolution lies in how well these components can communicate with each other, and work together in finding the most optimised solution for an assigned task. It is achieved by sharing data collected from a network of sensors. This data is communicated via images, videos, and a variety of other signals, attracting unwanted attention of hackers. The protection of such data is therefore pivotal, as is maintaining its integrity. To this end, this paper proposes a novel image watermarking scheme with potential applications in Industry 4.0. The strategy presented is multipurpose; one such purpose is authenticating the transmitted image, another is curtailing the illegal distribution of the image by providing copyright protection. To this end, two new watermarking methods are introduced, one of which is for embedding the robust watermark, and the other is related to the fragile watermark. The robust watermark's embedding is achieved in the frequency domain, wherein the frequency coefficients are selected using a novel mean-based coefficient selection procedure. Subsequently, the selected coefficients are manipulated in equal proportion to embed the robust watermark. The fragile watermark's embedding is achieved in the spatial domain, wherein self-generated fragile watermark(s) is embedded by directly altering the pixel bits of the host image. The effective combination of two domains results in a hybrid scheme and attains the vital balance between the watermarking requirements of imperceptibility, security and capacity. Moreover, in the case of tampering, the proposed scheme not only authenticates and provides copyright protection to images but can also detect tampering and localise the tampered regions. An extensive evaluation of the proposed scheme on typical images has proven its superiority over existing state-of-the-art methods.