Block-Adaptive Chaotic Watermarking with Enhanced Tamper Localization for Medical Image Authentication

Medical images have recently been exposed to unintentional and intentional destruction by unauthorized persons in the storage and transfer of medical images. This undermines the integrity and reliability of diagnostic information. Thus, the paper outlines a watermarking system to authenticate against medical image manipulation. The process involves the application of a sequential chaotic map to encode patient information on a QR code. A texture feature and entropy-based hybrid model are used to determine appropriate image regions to be used in embedding. The system also has an image manipulation detection and locating tool to maintain diagnostic content. X-ray, CT and MRI images were used for experimentation. The findings show reasonable visual quality and data recovery, and achieve high imperceptibility with PSNR value exceeding 56 dB across the different types of attacks, which proves that the proposed method is reliable in ensuring that medical images are not compromised to a range of attacks.