Robust Audio Watermarking using Hybrid lWT-DCTSVD Method and Arnold Scrambling

Digital watermarking is one solution to protect the copyright of digital data from illegal distribution, copyright infringement, and illegitimate sharing. Generally, audio watermarking is used to embed the owner or creator’s identity (watermark image) into the audio signal. The methods used in this paper are integer wavelet transform (IWT), discrete cosine transform (DCT), and singular value decomposition (SVD). First, the audio host signal is reshaped into the ۲D matrix to decompose by the ۲D-IWT to select the frequency sub-bands. The output of the ۲D-IWT will be transformed by the ۲D-DCT, where the domain conversion will be carried out, from the time domain to the frequency domain. Furthermore, the output signal in the form of a matrix from DCT will be processed by SVD to obtain UA, SA, and VA matrices. From the three matrices, the SA matrix is chosen to be inserted into the watermark. Before the embedding process, the watermark image is encrypted using the Arnold transform for further security, and then, the scrambled watermark image is also decomposed by ۲D-IWT, ۲DDCT, and SVD to obtain the SW matrix. After that using the scaling factor, SA and SW are combined to produce the SN matrix. Finally, the watermarked audio is obtained by applying the inverse SVD, inverse ۲D-DCT, and inverse ۲D-IWT to the output matrix SN and reshaping it to a ۱D signal. The simulation results show that the proposed watermarking method is very durable and strong against various types of attacks.

Keywords:

audio watermarking , integer wavelet transform , discrete cosine transform , singular value decomposition.