Study On Steganography In Source Camera Anti-forensics

Steganography as an important branch of information hiding, has been received more and more g attention of the researchers. This paper based on digital image hiding method to carry on the related research. The relationship between steganography and steganalysis likes spear and shield, which against each other and promote each other. We analyzed the advantages and disadvantages of steganography, and based on the high detection rate when the camera model is matching proposed the method of combining source camera anti-forensics and steganography to reduce the detection rate of steganalysis. This paper main research content is as follows:(1) This paper introduces five classic JPEG image steganographic algorithms, and analyzes their performance, then choses the nsF5 algorithm as the final steganography method, which has relative good property.(2)After analyzing the signature substitute method of source camera anti-forensics based on sensor pattern noise, according to the influence of parameters on the correlation between images and camera approximate linear change, we constructed two-dimensional linear fitting model and realized source camera ant-forensics. The experiment shows that anti-forensics method of this paper is effective.(3)This paper proposed the method of combining source camera anti-forensics and steganography, which improved steganographic security. We embedded information in the forgery images, and compared stegaographic security with original images. Firstly, under the source camera analysis model, compare the security of original images of source camera and forgery images. Secondly, under the under the target camera analysis model, compare the security of original images of target camera and forgery images. Experimental results confirm that both in the source camera model, or under the target camera model, the stegaographic security of forged image are relatively high. Although the method can improve security, the embedding rate of forgery images is very low. This paper improved the anti-forensics algorithm, and experimental results show that the optimized algorithm can effectively increase the biggest embedded rate. Compared with the original image, the stegaographic security of forgery image are greatly improved.