Application of Machine Learning Techniques for Generated Speech Detection
Segey Litovkin, Pavel Laptev, Evgeny Yurievich Kostyuchenko · 2024
This work is aimed at investigating the possibility of generated speech detection using machine learning techniques. The study was conducted on the basis of open datasets (CFAD, FoR), as well as on the basis of personal dataset created by the model of DeepFake speech generation RVC-v2. The study proved that speech could be detected on all three datasets as well as on the combination of all three datasets at 16 kHz sampling rate.