Robust Zero Watermarking Algorithm for Medical Images Based on BRISK and DCT
Fangchun Dong, Jingbing Li, Uzair Aslam Bhatti · 2023
As mobile internet technology continually advances, innovations in the realm of medical image processing are consistently evolving and improving. Within this field, digital watermarking technology plays a crucial role, offering robust security for medical image information. Considering the shortcomings of existing medical image watermarking methods when confronted with geometric attacks, this study introduces a robust zero-watermarking algorithm that employs the BRISK and discrete cosine transform (DCT) techniques to address the challenges associated with medical image watermarking. First, the BRISK algorithm is employed to extract the feature descriptor matrix from the medical image. Subsequently, the descriptor matrix undergoes a transformation using DCT, the perceptual hashing algorithm processes the low-frequency components to obtain a 32-bit binary eigenvector. Prior to embedding, the watermark information undergoes encryption using a logistic map algorithm., thereby increasing its security. the zero watermarking approach enables the algorithm to embed and extract watermarks without modifying the original medical image data. Experimental outcomes demonstrate that the proposed algorithm displays a certain level of robustness against both conventional and geometric attacks. It is particularly effective in mitigating geometric attacks. Compared with other methods, this algorithm has certain practical application values in the medical field.