Homologous Audio Copy-move Tampering Detection Method Based on Pitch

In order to verify the authenticity and reliability of the audio content, a detection method based on the pitch feature of homologous audio copy-move forgery is proposed. The principle of this method is that even if the same person speaks the same word twice, the pitch sequences of the two syllables are different. Firstly, the homologous audio should be non-linearly preprocessed. And then the pitch sequence is extracted as the feature of the audio. Because homologous audio focuses more on the characteristics of speech itself, and the sequence of pitches extracted from different syllables is usually completely different. Next, the difference between the individual syllables is calculated. Finally, it is determined whether the corresponding syllable has a copy-move forgery by comparing the difference between the syllables and the set threshold. The experimental results show that the method is robust to various conventional attacks. And the proposed algorithm can accurately locate the syllables which is copy-move forgery, and has better efficiency.