K-Means-Based Watermarking Algorithm For medical image
Mohammed Elhabib Kahla, Monir Beggas, Abdelkader Laouid · 2023
In recent years, multimedia and visual data management have grown substantially. The volume of this kind of data creates many challenges, including protecting multimedia content from misuse. One of the most various types of multimedia content is the image. The commonly used protecting image technique is watermarking, especially in medical images. Hence, improvements and modifications must be made to address the challenges associated with image watermark size and manipulations within multimedia data, including medical images. To this end, this paper proposes a medical image protection based on a watermark technique that employs the k-means algorithm. The proposed watermarking technique uses k-means to reduce the size of the watermark image and create a binary hash image, which is embedded into the pixels of the medical image using the least significant bit (LSB) method. Hereafter, the binary hash image is extracted from the watermarked image, and a bitwise XOR operation is applied between the extracted binary hash image and the origin-embedded binary hash image to detect occurred manipulations within the images. The PSNR has been used for evaluating the watermarking technique and subjecting the technique to various attacks involving different types of noise. The experiments and analyses show that the proposed watermarking technique can detect the manipulations within the images and reduce watermark image size, making it more suitable for secure medical images.