Detection Of Fake Audio

Abstract

Audio deep is the next frontier for company compromise scams, and it's becoming more usual for crooks to get access to corporate funds throughdeception. With the popularity and capability of audio deep fakes growing, it's more critical than ever to develop defences against deep fakes used for nefarious purposes. The goal is implemented using Python and the CNN technique.
The model is fed an image dataset of a frequency analysis audio sample. The model uses visual representations of audio clips, which are then fed into resnet34 and trainedto achieve accurate accuracy, to distinguish between real and fake audio. Recent breakthroughs in deep learning and other related fields.