Cross-Representation Transferability of Adversarial Attacks: From Spectrograms to Audio Waveforms
Abstract
This paper shows the susceptibility of spectrogram-based audio classifiers to adversarial attacks and the transferability of such attacks to audio waveforms. Some commonly used adversarial attacks to images have been applied to Mel-frequency and short-time Fourier transform spectrograms, and such perturbed spectrograms are able to fool a 2D convolutional neural network (CNN). Such attacks produce perturbed spectrograms that are visually imperceptible by humans. Furthermore, the audio waveforms reconstructed from the perturbed spectrograms are also able to fool a 1D CNN trained on the original audio. Experimental results on a dataset of western music have shown that the 2D CNN achieves up to 81.87% of mean accuracy on legitimate examples and such performance drops to 12.09% on adversarial examples. Likewise, the 1D CNN achieves up to 78.29% of mean accuracy on original audio samples and such performance drops to 27.91% on adversarial audio waveforms reconstructed from the perturbed spectrograms.
Keywords
Cite
@article{arxiv.1910.10106,
title = {Cross-Representation Transferability of Adversarial Attacks: From Spectrograms to Audio Waveforms},
author = {Karl Michel Koerich and Mohammad Esmaeilpour and Sajjad Abdoli and Alceu de Souza Britto and Alessandro Lameiras Koerich},
journal= {arXiv preprint arXiv:1910.10106},
year = {2020}
}
Comments
8 pages