Audio recapture detection using deep learning
Authors:
- 1 July 2015
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 478-482
Abstract
Since the audio recapture can be used to assist audio splicing, it is important to identify whether a suspected audio recording is recaptured or not. However, few works on such detection have been reported. In this paper, we propose an method to detect the recaptured audio based on deep learning and we investigate two deep learning techniques, i.e., neural network with dropout method and stack auto-encoders (SAE). The waveform samples of audio frame is directly used as the input for the deep neural network. The experimental results show that error rate around 7.5% can be achieved, which indicates that our proposed method can successfully discriminate recaptured audio and original audio.Keywords
This publication has 17 references indexed in Scilit:
- Spoofing and countermeasures for speaker verification: A surveySpeech Communication, 2015
- Deep Learning of Representations: Looking Forward2013
- Detecting Replay Attacks from Far-Field Recordings on Speaker Verification SystemsLecture Notes in Computer Science, 2011
- Revealing real quality of double compressed MP3 audio2010
- Defeating fake-quality MP32009
- Learning Deep Architectures for AIFoundations and Trends in Machine Learning, 2009
- Microphone Classification Using Fourier Coefficients2009
- Learning multiple layers of representationTrends in Cognitive Sciences, 2007
- Digital audio forensics2007
- Digital audio recording analysis: the Electric Network Frequency (ENF) CriterionThe International Journal of Speech, Language and the Law, 2005