基于特征谱图融合的音频信号复制粘贴篡改检测算法OA

Copy-and-paste tampering is characterized by its manipulability and imperceptibility.Mean-while,the existence of various post-processing operations makes audio tampering detection more difficult.For audio copy-paste tampering with post-processing,the paper proposes an audio copy-paste tampering de-tection algorithm based on the fusion of feature spectrograms.Firstly,the audio signal is converted into Hil-bert spectrogram and Mel spectrogram respectively,and the two feature spectrogram images are weighted and fused to obtain a more robust feature spectrogram image.Further the key points are extracted from the fused feature image by utilizing scale invariant feature transform and matching the keypoints.Then the du-plicate regions on the image are identified with an agglomerative hierarchical clustering algorithm to imple-ment audio copy-paste tampering detection.The experimental results show that the proposed method is highly robust against the various common post-processing attacks.Compared with previous state-of-the-art audio signal copy-paste tampering detection methods,the proposed method in this paper has higher detec-tion precision,recall and robustness.