Modify the Quantization Table in the JPEG Header File for Forensics and Anti-forensics

Modify the Quantization Table in the JPEG Header File for Forensics and Anti-forensics

연구 분야: Cryptography

학회: International Workshop on Digital Watermarking

As Joint Photographic Experts Group (JPEG) compression is widely used in image processing and computer vision, the detection of JPEG forgery has become an important issue in digital image forensics, and many related works have been reported. However, these works all assume that the quantization table in the JPEG header file is real, and corresponding research is carried out based on this assumption. This assumption leaves a potential flaw for those wise forgers to confuse or even invalidate the current JPEG forensics detectors. Taking double JPEG compression forensics as an example, if the quantization table in the header file is modified, it will cause the algorithm to fail. However, the tampering of header files not only brings negative effects, it can also improve the forensics performance of the algorithm. According to our analysis and experiments, increasing the step in the quantization table in the header file can lead to the failure of the forensics algorithm, and reducing the step in the quantization table in the header file can improve the performance of the existing algorithm. Based on this observation, we propose a general forensics and anti-forensics model by replacing the quantization table in the header file. The experimental results on the UCID database show that the scheme is effective for obfuscating and improving the three typical double JPEG compression forensics work.