作者: Xiaochuan Chen , Yunhong Wang , Tieniu Tan , Lei Guo
关键词: Histogram 、 Information hiding 、 Support vector machine 、 Computer science 、 Steganalysis 、 Embedding 、 Classifier (UML) 、 Artificial intelligence 、 Feature extraction 、 Data mining 、 Contextual image classification 、 Pattern recognition
摘要: In this paper, a novel steganalysis method based on statistical analysis of empirical matrix (EM) is proposed to detect the presence hidden message in an image. The projection histogram (PH) EM used extract features composed two parts: moments PH and characteristic function PH. Also, extracted from prediction-error image by Yun Q. Shi et al, (2005) are included enhance performance. SVM utilized as classifier. A test database constructed, which detailed for different categories comparison with methods prior arts conducted. Experiments show that we more effective than our could blindly data hiding various embedding schemes high performance