CRT-based fragile self-recovery watermarking scheme for image authentication and recovery

Abstract

Fragile watermarking is one of the effective techniques for authentication of digital documents and images. However, recovering the content of the tampered region in a watermarked image is a challenging task while considering conflicting criteria of imperceptibility and watermark embedding capacity. In this paper we propose a Chinese remainder theorem (CRT)-based watermarking scheme which can recover the original contents in the tampered region of the digital content while maintaining imperceptibility criterion. High peak signal to noise ratio (PSNR) and large watermark capacity can be achieved by using the CRT-based embedding scheme. Since only modular operations are involved in computation of the CRT-based technique, it provides computational advantage as it involves only modular arithmetic. Besides, CRT-based technique introduces additional security to the watermarking scheme. By taking several digital images, we have shown that the proposed technique can recover the tampered contents effectively. We have also considered forgery detection on a digital cheque, eCheque, and shown that the proposed technique can detect and recover the original content from the forged cheque.