High imperceptibility medical image watermarking scheme based on slantlet transform by using dynamic visibility threshold

• Main Authors: Tan, Chi Wee, Mohd. Rahim, Mohd. Shafry, Tan, Gloria Jennis, Sulong, Ghazali, Chaw, Jun Kit
• Format: Conference or Workshop Item
• Published: 2021
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://eprints.utm.my/id/eprint/94340/; http://dx.doi.org/10.1109/ICIDM51048.2020.9339651
• Institution: Universiti Teknologi Malaysia

Digital watermarking is being used for increasing the level of effectiveness and security of the documents transfer over the internet, especially in the biomedical field. This study examined and explored the new working domain for medical image watermarking by introducing a Dynamic Visibility Threshold (DVT) as a crossover from pure mathematics over the traditional fixed visibility threshold. Initially, for providing higher robustness against the watermarking attacks, the watermarking process is done on Slantlet Transform, an improved version of Discrete Wavelet Transform. By investigating the root of the medical image properties, we introduce the descriptive statistics namely Quantile Theory to deliver a dynamic value from the biomedical itself for the embedding process. Later, we had identified that the Third Quartile is the most optimum value for the DVT in terms of higher imperceptibility. This paper uses a standard dataset from BRAINIX, a medical database provided by OsiriX, Pixmeo SARL, and the effectiveness of the proposed method is analysed using MATLAB R2013a. The most obvious finding to emerge from this study is the convergence of descriptive statistics from pure mathematics and medical image watermarking from computer science, therefore shine new light on dynamic visibility threshold through mathematical theories. With the presence of descriptive statistics, quartile theory is being used to increase the imperceptibility of the watermarking in biomedical images.