A Novel Semi-Fragile Image Watermarking, Authentication and Self-Restoration Technique Using the Slant Transform

Abstract

A novel semi-fragile digital watermarking method based on the slant transform (SLT) for image authentication and self-restoration is introduced in this paper. The watermark bits are embedded into the middle frequency region of each block after applying SLT of the original image. The original image is further compressed and then embedded into the least significant bits (LSBs) of the watermarked image for subsequent self-restoration. The tampered regions of the watermarked image can be detected and localised by extracting the embedded watermark to compare with the original watermark for authentication. Localised tampered regions are self-recovered by extracting the LSBs of the watermarked image. Results achieved show that the SLT algorithm is more robust, faster and accurate than other transform methods based on the DCT and pinned sine transform. It could survive cut-and-paste attacks with robustness to JPEG attacks up to QF=75 compression prior to image authentication. The tampered regions can be detected with an average 91% detection rate after surviving the compression attack.