Local feature based self-embedding fragile watermarking scheme for tampered detection and recovery utilizing AMBTC with fuzzy logic

Abstract

The aim of designing self-embedding fragile watermarking is copy-right protection, tamper localization and image recovery. This paper proposes a self-embedding fragile watermarking scheme for tamper detection and recovery based on local image characteristics using Absolute Moment Block Truncation Coding (AMBTC) and fuzzy logic. The original image is divided into (4×4) non-overlapping blocks. The watermark of each block are generated from 6 MSBs of each pixels of the block. Using similarity matrix, the blocks are categorized into smooth and complex depending on similarity among neighboring pixels. In smooth block, mean is used as recovery data for their high similarity existence whereas complex block uses AMBTC image as it has a compressed representation of an image and preserves the features of the cover image. Recovery data are embedded into 2 LSBs of corresponding mapped block. The efficiency of the proposed scheme is analyzed in terms of Tamper Detection Rate (TDR) and Tamper Recovery Rate (TRR). The performance of the proposed scheme is analyzed against different attacks applied on different regions of the watermarked image. The experiment results show that the imperceptibility and visual quality of the watermarked image as well as recovered image is satisfactorily high even at 50% tampering rate.