Detection of impostor and tampered segments in audio by using an intelligent system

The transmission of audio data via the Internet Things makes such vulnerable to tampering. Moreover, availability sophisticated tampering tools has allowed mobsters change context by altering their segments. Tampered may result in unpleasant situations for any member society. To avoid circumstances, a new forgery detection system is proposed this study. This can be deployed edge devices identify impostors and data. implemented using state-of-the-art mel-frequency cepstral coefficient features. Meanwhile, Gaussian mixture model used train validate system. evaluate system, dataset tampered audios created mixing recordings from two different speakers. performance authenticating genuine between 92.50% 100%, that detecting forged 99.90 100%.