Image Fragile Watermarking With Two Authentication Components for Tamper Detection and Recovery

Abstract

This study developed a new fragile watermarking method that utilizes the image feature as a watermark component to be inserted into the image itself. There are two watermark components, for the process of tamper detection and recovery. In this scheme, the original image is split into multiple blocks of uniform size and Iookup tables for random blocks are generated using the 1-D transformation algorithm. In the process of watermark generation for tamper detection, we used parity bit for 5 MSB and the average block for 2×2 pixel image, this last process resulted in two check bits generated by random key with a certain size. The watermark is then inserted into three least significant bits, where the watermark used authentication is inserted in the original pixel, while for bit recovery is inserted into the appropriate mapping blocks. In decoding, feature information as a watermark is reconstructed and the watermark image is recovered based on reconstructed data along with some other post-processing methods. When compared with the previous method indicates that this algorithm can improve image quality.