An SVD-based audio watermarking using variable embedding strength and exponential-log operations

This paper presents a singular value decomposition (SVD) based audio watermarking scheme using variable embedding strength, exponential and logarithm operations for copyright protection. In our proposed scheme, initially the original audio is segmented into non-overlapping frames. Discrete cosine transformation (DCT) is applied to each frame. DCT coefficients of each frame are divided into sub bands and energy of each sub band is calculated. Exponential operation is performed on the largest energy sub band of the DCT coefficients of each frame. SVD is applied to the exponential coefficients of each of the largest energy sub bands represented in matrix form. The mean of the singular values of each of these matrices are then calculated. Watermark information is embedded into the largest singular value by using a variable embedding strength. Simulation results indicate that the proposed watermarking scheme shows strong robustness against different attacks and it provides low error probability rates. Moreover, it outperforms the state-of-the-art watermarking methods in terms of imperceptibility, robustness, and data payload.