An optimized hierarchical encryption technique for tamper recognition

Abstract

Digital images contain sensitive information that needs to be watermarked for ownership authentication and copyright verification. It is crucial to have a method to detect and recognize tampering when images are sent over insecure channels. The watermarking scenario becomes more complicated when an intruder is precluded from obtaining a watermark signal from a watermarked image since it can expose future point-to-point correspondences. The proposed scheme utilizes a hierarchical strategy for improving the security of the semi-fragile watermarking scheme that requires fewer data to be exchanged before each transaction. Consequently, the proposed watermarking method obtains a trade-off between robustness and imperceptibility by using a meta-heuristic approach, namely, the Sine Cosine Algorithm (SCA). Furthermore, an Artificial Neural Network (ANN) is built using the Softmax classifier to recognize possible attacks that might be performed by an intruder. The whole scheme is presented in the form of a GUI with the attack recognition triggered from the receiver’s side. Experimental results show improved image quality metrics like PSNR, correlation coefficient, and structural similarity when the scaling factor used in the watermarking algorithm is optimized using SCA.