NOISPROOF DIGITIAL AUDIO WATERMARKING IN MIMO AUDIO STEGOSYSTEMS
M. V. Gofman
Emperor Alexander I St. Petersburg State Transport University
Annotation: The article presents methods of noiseproof combined digital audio watermarking, considering the peculiarities of audio stegosystems using channels with multiple (spatial) input and multiple (spatial) output. Such audio stegosystems include several acoustic speakers (transmitters)
and several microphones (receivers), which makes it possible to embed a watermark in the parameters of digital audio signals in the spatial, frequency and time domains. The methods for creating, embedding, and extracting a combined watermark developed in the article are aimed at ensuring a high degree of noiseproof when transmitting a watermarked audio signal through an airborne audio channel at low embedding forces. The method of creating and embedding the combined watermark is based on the transformation of identifying digital information using a synthesized steganographic key into a special sequence and subsequent space-time-frequency coding of the digital audio signal parameters by it. The main attention is paid to the development of a method for detecting and extracting an information bit (extracting) a combined watermark by an authorized receiver in an audio signal transmitted over an airborne audio channel with noise. In this case, it is assumed that the authorized receiver does not know the parameters of the digital audio signal that have been watermarked (blind reception). In the proposed method for extracting a watermark, the decision on its detection is made according to the threshold principle based on the estimation of the peak factor of the kurtosis of the target random variable.
Keywords: Audio Stegosystem, MIMO, Channel with Multiple Input and Multiple Output, Watermarking, Digital Audio Signal, Combined Watermark, Airborne Audio Channel, Noiseproof.
Pages 83-95