A New Method for Blind Image Steganalysis
Yang Yi-xian · Beijing Youdian Xueyuan xuebao · 2006
Image quality metrics have two inputs: the original image and the stego-image.Whereas blind image steganalysis only has one input: the stego-image.In order to fit blind image steganalysis,the image quality metrics must be altered: the original images are replaced by white images.Using four image quality metrics,the distance between stego-images and white images can be obtained.Regarding the distance as eigenvectors and using SVM(support vector machine) to classify them into two classes: one includes hidden information,the other one doesn't.Band-pass filter property of human visual system is used to improve correct detection rate.Experiments show that this method is effective to nature images.