An Adaptive Image Watermarking based on Bellman-Ford Algorithm
Brahim Ferik, Lakhdar Laimeche, Abdallah Meraoumia, Abdelkader Laouid, Muath AlShaikh, Khaled Chait · 2023
Ensuring the integrity and authenticity of images is a paramount task in the digital multimedia context. The emergence of digital technologies necessitates the development of robust techniques to safeguard medical images against tampering or unauthorized alterations. One such crucial approach is watermarking, which involves embedding imperceptible and secure data within the images to facilitate subsequent verification and authentication. This paper proposes a novel medical image watermarking technique utilizing fingerprint data and a hybrid standard deviation-Bellman Ford algorithm for enhanced security and minimal distortion. The host image is divided into blocks, and the standard deviation quantifies texture variations. The Bellman-Ford algorithm then identifies optimal paths in the image graph corresponding to regions suitable for embedding. Our methodology diverges from existing techniques by enabling adaptive watermark embedding within textured domains while circumventing smooth regions. This study concentrates on the targeted objects within the image to uphold their integrity by inserting the adaptive watermark in the regions of interest. Extensive evaluation of various modalities proves the efficacy of the proposed method with an average Peak Signal Noise Ratio of 60.67dB and Structural Similarity Index of 0.9999, demonstrating marked improvements in imperceptibility over state-of-the-art techniques. The approach is efficient, with embedding and extraction times of 0.0094s and 0.0083s, respectively.