Research On Passive Forensics Of JPEG Image

With the popularity of the Internet and the rapid development of digital media, digital image has been widely used and became an important information carrier. In the digital era, with the emerge of the image processing softwares, it is possible to modify the digital images simply and rapidly. In addition, the doctored image without leaving evidence of tampering. It is hard to distinguish the authenticity of those images by visual inspection, therefore, it is need some reliable and robust detection methods to do this.At present, the digital images used on the internet and the output images of the photographic equipment are stored in JPEG format. In addition, the tampered images are always stored in JPEG format. Hence, the research of the JPEG image passive forensics has important academic meaning and practical value. Based on this, the paper launches the research of the JPEG image passive forensics. The main contents of the paper are as follows:(1) Based on the quantization coherence of the double quantization process in the forged process, a project of JPEG image forensic based on quantization coherence is proposed. Firstly, the scheme by compressing the JPEG image twice in order. Then, calculating the quantization coherence introduced by the JPEG compression. Last, because the quantization coherence of the background areas and the tampered region is difference, so detect the JPEG image by analysing the quantization coherence between them. The simulation results indicate that the approach is effective in detecting the doctored JPEG images and localizing the tampered area.(2) By researching on the JPEG compression process, when using different quality factor in the image compression process will produce different degree of distortion. Based on this, the paper proposes an image detecting scheme. On the first, the scheme by estimating the original quality factor of the JPEG image. Then, compress the JPEG image by the original quality factor and analysis the distortion of the compress process. Because of the distortion of the image pixels between the background areas and the tampered region is different, hence, the doctored region can be distinguished from the background areas. The simulation results show that the approach is effective in detecting the doctored JPEG images and localizing the tampered areas.