Secure and Effective Image Integrity and Copyright Protection Using Two-Layer Authentication with Integer Wavelet Transform
Year: 2025
Type: article
Abstract: The pervasive adoption of internet and digital communication technologies renders multimedia data, particularly images, vulnerable to piracy due to the ease with which unauthorized users can manipulate them. This research advocates for the application of IWT and SVD grounded in human visual characteristics to embed watermarks into host images, thereby fortifying their resilience against attacks while minimizing distortions in image quality. Authentication bits extracted from the content of 3×3-pixel blocks within the image are subsequently incorporated into the watermarked image, positioned within the LSB of each pixel to facilitate the authentication of RGB images against tampering attacks. The efficacy of the proposed method is evaluated through various geometric, compression, image processing, and tampering attacks, alongside comparison to alternative dual watermarking systems. Experimental findings reveal the precise identification of tampering in most instances, even when alterations impact only a small number of blocks. Moreover, the method demonstrates substantial robustness, even when subjected to extensive image processing attacks, with a PSNR value of around 48 dB and an authentication accuracy of 0.98 for tampering attacks, thus indicating potential for robust copyright protection.
Authors: Ng Ching Onn, Siti Nur Avivah, Mohammad Minhazul Alam, Abubakar Ahmad
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Funder: Universiti Malaysia Pahang
Award: G8758726568
OpenAlex ID: W4409512263