A Novel Calibration Used for Blind Steganalysis of JPEG Images
Ling Wang, Jing Xin Hong, Chen Guang Xu, Qian Chen, Yi Xiong Zhang · Advanced materials research · 2014
In this paper, we present a new calibration technique aimed at blind steganalysis for JPEG images, which can magnify the difference between cover images and stego images. The calibration can be considered as a preprocessing of stego image before extracting the features. So the calibrated features are calculated as the difference between a specific function calculated from the original stego image and the same function obtained from calibrated version. Moreover, the calibrated feature was used to train SVM (support vector machine), a nonlinear classifier, which is effective in class separation. For comparison, a database composed of 6690 cover and stego images (generated by using four different embedding schemes) was established. Based on this database, we conducted extensive experiments and drawn a conclusion that the steganalysis based on our novel calibration can detect the stego images with high accuracy.