Neural network based audio watermarking algorithm

Longhan Cao, Xiao Wang, Zhujing Wang, Sen BaiProceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE 2005 · DOI

Abstract

A novel audio watermarking algorithm based on neural networks was proposed. By transforming original audio sequence into 1D wavelet domain and selecting proper positions, several watermark bits were embed. Before transmitting, it effectively utilizes neural networks to learn the relation characteristics between original audio and watermarked audio. Due to the learning and adaptive capabilities of neural networks possessing, the trained neural networks almost exactly extract the watermark from the watermarked audio against audio processing attacks. Extensive experimental results show that the proposed method significantly possesses robustness. It is immune against such attacks as low pass filtering, addition of noise, resampling and medium filtering.
BibTeX
@inproceedings{cao2005neural,
  author = {Longhan Cao and Xiao Wang and Zhujing Wang and Sen Bai},
  title = {{Neural network based audio watermarking algorithm}},
  booktitle = {Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE},
  year = {2005},
  doi = {10.1117/12.664309},
}

Loading related content...

Figure