A robust fragile watermarking approach for image tampering detection and restoration utilizing hybrid transforms
S. Prasanth Vaidya, Kandala N. V. P. S. Rajesh, Chandra Mouli, Hatim Ghazi Zaini, Amar Y. Jaffar, Prabhu Paramasivam, Sherif S. M. Ghoneim
Abstract
This study presents a novel fragile watermarking technique to detect and restore image tampering, enhancing security in digital image transmission. The proposed method integrates Schur decomposition and discrete wavelet transform (DWT) for watermark embedding, ensuring robustness against attacks compared to existing methods. Schur decomposition provides numerical stability in matrix factorization, while DWT enhances resilience through multi-resolution analysis. A semi-blind extraction algorithm, relying only on a secret key, enables active tampering detection without requiring the original image. Upon detection of distortions, the proposed recovery mechanism restores the tampered regions of the image. The effectiveness of the proposed scheme is validated through structural similarity, peak signal-to-noise ratio, and normalized cross-correlation metrics, demonstrating superior performance compared to existing methods. This approach is applicable to secure medical imaging, forensic investigations, and copyright protection, ensuring image integrity in real-world scenarios.