An Algorithm Of Resist Cropping Robust Audio Watermark Based On Wavelet Transformation

This paper proposes a robust audio watermark algorithm based on discrete wavelet transform (DWT) to resist cropping attack. The approach initially transforms the audio signal into wavelet domain by DWT, aiming at extracting the low frequency coefficients. A binary image via Arnold preprocessing embedded in audio low frequency coefficients on wavelet domain through the way of quantization steps. This algorithm extract and detect the watermark did not need the original watermark. Simulation experiments show that the performance of the method is ideal under various signal processing manipulations such as low pass filtering, overlay noise, re-sampling, re-quantization, random cropping, dithering, MP3 compression, especially robust to cropping attacks.