A Block-Level Image Tamper Detection Scheme Using Modulus Function Based Fragile Watermarking

Intentional illegitimate modification in any digital image data is one of the general malversations in the existing digital domain. Therefore in this work, the authors have devised a fragile watermarking technique for the localization of illegitimate modifications in the digital image content effectually. The proposed technique detects forged digital image content strongly in the block-level of two successive pixels. This scheme is performed at the block level, where the actual image is partitioned into non-overlapping blocks of size (1 x 2) pixels. Afterward, the block-level authentication code is generated from the MSB of two successive pixels of each block using averaging and modulus operations. The generated authentication code/watermark is encrypted using a logistic-map-based chaotic key series. The encrypted authentication code is embedded into two successive pixels of the corresponding block. Further, any form of forgery in the watermarked image can be detected by comparing its extracted authentication code and regenerated authentication code. The proposed procedure is successfully experienced on a variety of grayscale images, and the experimental results exhibit that watermarked images generated by this scheme are of considerably high quality in terms of PSNR, IF, and SSIM. The proposed method is capable of efficient forgery detection by achieving very high accuracy, NC, and true positive rate while maintaining considerably low false-positive and false-negative rates.