Steganalysis using difference images and bit-plane splitting

Steganalysis, in contrast to steganography, is a method that focuses on detecting secret messages hidden in carrier media. In this paper, a blind steganalysis method for digital images is proposed. We describe the statistical characteristics of the difference image, and the statistical characteristics of the images in bit-plane to exploit features. Then we use the multi-layer perceptron (MLP) to classify images. Extensive experimental results show that the proposed method is efficient than some previous methods.