Digital audio reversible watermarking algorithm based on side information prediction and histogram shifting

Abstract

The invention discloses a digital audio reversible watermarking algorithm based on side information prediction and histogram shifting. The algorithm includes the following steps that the prediction coefficient of audio data and the prediction value of a sample are calculated; the prediction value is subtracted from the sample value to obtain a prediction error, two different histograms are constructed according to two modes of the side information prediction algorithm, and then watermarking information is embedded in the prediction error to obtain the prediction error containing a watermark; the prediction error containing the watermark is embedded in a next sample value of the current sample to obtain an audio value containing the watermark; an audio signal containing the watermark is received at the receiving end, the prediction value is subtracted from the sample value, an embedded signal is extracted by means of prediction error histogram shifting characteristics, and an original audio signal is restored in a lossless mode. Compared with an original prediction algorithm, the algorithm improves prediction performance; non-causality and histogram shifting are combined, so the algorithm has good embedability and low distortion rate when the embedding rate is low.

Description

technical field [0001] The invention relates to the field of audio data processing, in particular to a digital audio reversible watermark algorithm based on side information prediction and histogram translation. Background technique [0002] Audio reversible watermarking is an important branch of multimedia information security and hiding. It has two main characteristics: the first is that the watermark can authenticate the integrity and authenticity of the audio, and locate tampering; the second is that the carrier can be restored without damage, and That is to say, the watermarking algorithm is reversible. As a processing method of audio data, reversible watermarking mainly embeds hidden information into the audio, then extracts the hidden information at the receiving end, and finally restores the audio value without loss. [0003] The criteria for judging the quality of the reversible watermark embedding algorithm are mainly measured from several aspects: the embedding c...

AI Technical Summary

Problems solved by technology

[0014] Disadvantages: Forecasting before and after is to set the prediction coefficient to 0.5, and the average value of the two sample values ​​before and after the sample is obtained to predict, and the predicted value is not very accurate.

This algorithm takes advantage of the auditory characteristics of the human ear. After embedding information, the original audio data cannot be recovered, that is, this algorithm is not a digital audio reversible watermarking algorithm, and it is not suitable for some occasions that require high audio quality.

Images