Detection of Copy-Move Forgery in Audio Signal with Mel Frequency and Delta-Mel Frequency Kepstrum Coefficients

Digital multimedia security has taken a very important position with the developing technology. Detecting forgeries in audio signals is one of the most challenging application in the field of audio forensics. In this study, Mel Frequency and Delta Mel Frequency based methods are proposed to detect copypaste forgery in audio signals. Firstly, the proposed study YAAPT method was used as a pitch tracking algorithm to detect voiced/unvoiced segments from audio signals. Mel Frequency Kepstrum Coefficients (MFCC) and Delta-Mel Frequency Kepstrum Coefficients (Delta-MFCC) were used to extract features from the audio signal seperated into voiced/unvoiced segments. By averaging the obtained coefficients, 13 MFCC and 13 Delta-MFCC coefficients were obtained for each region. The Pearson correlation coefficient was then calculated to find the similarity between the segments. In the final stage, a threshold value was determined based on the obtained correlation matrix and it was determined whether copy-move forgeries were made in the audio signals.