Zero-Watermarking for Medical Images Based on Regions of Interest Detection using K-Means Clustering and Discrete Fourier Transform

Watermarking schemes ensure digital image security and copyright protection to prevent unauthorized distribution. Zero-watermarking methods do not modify the image. This characteristic is a requirement in some tasks that need integrity, such as medical images. obtain specific features for master share construction protect paper proposed zero-watermarking scheme based on K-means clustering ROI detection features. The algorithm classifies data according proximity of generated clusters. applied segmentation identify detect areas contain important information from Therefore, Discrete Fourier Transform (DFT) features, using high frequencies increase its robustness against geometric attacks. In addition, an edge Sobel operator QR code creation. type watermark avoids errors increases system. creation XOR logic operation between extracted selected watermark. method focuses despite it being tampered with. Many focus advanced processing experiments demonstrate presented robust signal-processing DFT coefficients efficiency robustness.