Audio Watermarking for Security and Non-Security Applications

Abstract

The digitization of audiovisual data is significantly increasing. Thus, to guarantee the protection of the intellectual properties of this digital content, watermarking has appeared as a solution. Watermarking can be used in reality in several types of applications that target two different contexts: the first for security applications and the second for non-security ones. In this paper, we carry a big interest in studying these two types of applications. Moreover, we propose a first digital watermarking scheme for security copyright protection applications, where we have involved neural network architecture in the insertion and detection processes, and integrated some masking phenomena of the human psychoacoustic model with linear predictive coding spectral envelope estimation of the audio file. Experiments proved the efficiency of exploiting perceptual masking with spectral envelope consideration in terms of imperceptibility and robustness results. In addition, we suggest a second audio watermarking technique for non-security content characterization applications based on a deep learning classification architecture. In this scheme, the extracted watermark advises about the audio class: music or speech, speaker gender, and emotion. The reported results indicated that the suggested scheme achieved a higher performance at the classification level, as well as at the watermarking properties.