Research On Audio Watermarking Scheme Based On Two-level Singular Value Decomposition And Grey Wolf Optimizer

The rapid development of the internet and multimedia technology has made it easier and more convenient to store,distribute and copy audio files,but it has also brought problems such as copyright infringement and malicious tampering with audio information.Audio watermarking technology is an effective method to solve the above problems.Audio watermarking technology uses redundancy of audio signal and masking effect of human auditory system to embed the watermark into an audio signal without affecting the audible quality of the host audio,so as to achieve the purpose of copyright protection and integrity verification of the audio.In the field of audio watermarking,the embedding position and embedding strength of the watermark have an important influence on the imperceptibility,robustness and capacity of the audio watermarking algorithm.In order to improve the performance of the audio watermarking algorithm,this paper focuses on the watermark embedding position and embedding strength based on two-level singular value decomposition(SVD)and grey wolf optimizer(GWO).The specific work is as follows:(1)To address the problem of embedding the watermark image in the host audio,a grayscale image audio watermarking algorithm based on Discrete wavelet transform(DWT)and two-level SVD is proposed.The algorithm firstly divided the original audio signal into several frames,then applied DWT to each frame and the low-frequency components are selected.The low-frequency components are decomposed by SVD operation to obtain the singular values.The maximum singular value of each frame is collected into a new matrix,and then the second SVD operation will be performed on the new matrix.The singular value of the watermark image after performing two-level SVD,is embedded into the singular value of the new matrix.The watermark image is embedded into the singular value after two-level SVD,which improves the capacity and robustness of the algorithm.In order to protect the security of the watermark information,the grayscale image is encrypted by using phase-truncated Fresnel transform(PT-Fr T)based optical asymmetric cryptograph.Experimental results show that the imperceptibility of the proposed algorithm is satisfied,and that high capacity and strong robustness is obtained.(2)To address the problem of watermark embedding strength,a grayscale image audio watermarking algorithm based on two-level SVD and GWO is proposed.Considering the imperceptibility,robustness and capacity of the watermarking algorithm,a new objective function is designed and using GWO optimize the objective function.The comparison with other objective function shows that the proposed objective function can better balance imperceptibility and robustness while ensuring high embedding capacity.(3)To address the problem of embedding watermark in a single form,a new high imperceptibility and robust audio dual-watermarking scheme based on synergy fuzzy-c means(FCM)clustering and GWO is proposed to embed two watermarks(a grayscale and a binary)into the audio signal.In the proposed scheme,the original audio signal is first split into several frames,and then these frames are divided into high-energy(HE)frames suitable for embedding and low-energy frames not suitable for embedding by using FCM clustering.Then the two-level DWT is performed on the HE frames,the grayscale watermark and binary watermark are respectively embedded into the low-frequency and high-frequency components by using different embedding methods.By designing a new fitness function and using GWO,we obtain the optimal two watermark strengths in order to balance the requirements of imperceptibility,robustness,and capacity.The comparison analysis shows that the proposed scheme has higher imperceptibility and provides superior performance.