An Audio Zero-Watermarking Algorithm Based on Wavelet and Cepstrum Coefficients Mean Comparison

Teng Hu,Ru Zhang,Jian Yi Liu
DOI: https://doi.org/10.4028/www.scientific.net/amr.811.508
2013-01-01
Advanced Materials Research
Abstract:This paper presented an audio zero-watermarking algorithm based on wavelet and cepstrum coefficients mean comparison. The Algorithm first divided the original audio into four parts, and then used the wavelet transform to conduct 3-level wavelet decomposition in every audio signal, extracting the approximate coefficients of the third layer wavelet to conduct cepstrum transform according to fixed length segment,and removing the part with high fluctuation at both ends of the cepstral coefficients, selecting the stable middle part to construct the watermark based on the two neighboring mean value relation. The simulation results showed that the algorithm had fine robustness under common attacks like resampling, requantization, MP3 compression, Gaussian white noise, low-pass filtering,wavelet denoising and echo attack.
What problem does this paper attempt to address?