Improved Robust, Invisible Digital Image Watermarking Using Discrete Cosine Transform and PM

Abstract

With technological advancements, the internet has become the primary medium for transferring information across the globe. This medium enables quick, easy, and accurate data transmission. However, transmitting information over the internet also presents significant security threats. In this context, digital watermarking has emerged as a promising technique for ensuring data authentication and integrity. In the proposed approach, the source image is processed in row-major order, and the Discrete Cosine Transform (DCT) is applied to generate low and high-frequency components. One bit of the authenticating image is embedded into each transformed DCT coefficient, excluding the first coefficient, to maintain low perceptual distortion and enhance PSNR. Early work on digital watermarking, including analyses of various schemes and their applications, is discussed to provide background. Unlike many existing methods that emphasize content authentication over strict integrity, this paper introduces a binary value-shifting technique based on the statistical characteristics of image pixels. This method embeds watermark data into the cover image to improve PSNR and enhance image protection. The implementation is carried out on the Windows platform using MATLAB. The proposed technique aims to secure digital images while achieving improved PSNR values. Digital watermarks are also used to verify the reliability and integrity of the signal or image data. The effectiveness of the approach is evaluated using normalized correlation to assess robustness against various attacks.