Voice Anti-Spoofing Protection
An intelligent system for detecting fake voices that helps counter threats such as phone fraud, falsified audio recordings, and cybercrime. The service is already deployed on the ML Portal of the University.
The system supports online detection, batch processing of audio files, and visual representation in the form of spectral heatmaps highlighting anomalous regions. This makes it valuable for forensic analysis, risk management systems, and educational platforms.
Based on artificial intelligence, the system identifies subtle indicators of synthetic speech (TTS, VC) and delivers interpretable results suitable for legal analysis. Special attention is given to privacy: the service does not analyze speech content, operates in a secure environment, and ensures action auditing.
