Artificial intelligence

Abstract: The approach to steganalysis of QIM steganography is considered. It is based on feature vectors of audio signals derived from co-occurrence matrix in time domain, which is sensitive to data embedding process. It also uses the binary classifier that implements the support vector machine and is used for separation of empty and stego containers. The efficiency of this approach wis investigated with using of various parameters of stego transformation.

Keywords: information security, steganography, steganalysis, audio signals, co-occurrence matrix, SVM

References: