A review on blind detection for image steganography

Blind steganalysis techniques detect the existence of secret messages embedded in digital media when the steganography embedding algorithm is unknown. This paper presents a survey of blind steganalysis methods for digital images. First, a principle framework is described for image blind steganalysis, which includes four parts: image pretreatment, feature extraction, classifier selection and design, and classification. We then classify the existing blind detection methods into two categories according to the development of feature extraction and classifier design. For the first category, we survey the principles of six kinds of typical feature extraction methods, describe briefly the algorithms of features extraction of these methods, and compare the performances of some typical feature extraction algorithms by employing the Bhattacharyya distance. For the second category, the development of classifier design, we make a survey on various classification algorithms used in existing blind detection methods, and detail the algorithms behind several classifiers based on multivariate regression analysis, OC-SVM, ANN, CIS and Hyper-geometric structure. Finally, some open problems in this field are discussed, and some interesting directions that may be worth researching in the future are indicated.