Improved Sine-Cosine Algorithm For Anti Forensics JPEG Compression

Abstract

In the past few years, the extensive appropriate of the JPEG image in transfer, storage, etc are rising. In the image, to recognize the local tampering the lossy character of the JPEG compression leaves of traces that are done using forensic agents. Here, by the JPEG compression, a difficult anti-forensic approach is developed to get rid of the traces left in both the spatial domain and DCT. A new improved sine-cosine (ISC) technique is developed on presented compression of anti-forensic format to optimally evaluate the noise-like signal to be augmented in the de-calibration and de-blocking operation. Additionally, in the optimization algorithm, a novel fitness function named Histogram Deviation (HD) is devised to appropriately balance the quality of visual and the undetectability of forensic. The simulation of the proposed anti-forensic compression method is done from the UCID database with the uncompressed images. The evaluation of the proposed technique is analyzed with the conventional techniques exploiting MSE, PSNR, and accuracy of classification as metrics. The simulation results show potential consequences such as high accuracy and low MSE that demonstrate the effectiveness of the proposed technique is deceptive the forensic agents.