Proceedings of the 5th ACM workshop on Digital rights management
Self-synchronization Blind Audio Watermarking Based on Feature Extraction and Subsampling
ISNN '07 Proceedings of the 4th international symposium on Neural Networks: Part II--Advances in Neural Networks
| Hi-index | 0.00 |
This paper proposes an audio watermark extraction technique which adopts Independent Component Analysis (ICA) for blind watermark decoding. Unlike the existing work, our method allows to combine data synchronization and watermark decoding into one optimization procedure, thus robust to transmission over different channels. Watermark encoder is designed as a nonlinear data embedding machine which is compatible to MPEG Layer1 Model 1. It is shown that the proposed ICA based watermark decoding scheme allows decoding watermark info accurately even though the watermark to signal ratio is less than -20 dB. The method is robust to stereo-to-mono conversions and performs very well when the channel noise level is high.