detecting-deepfake-audio-in-vishing-attacks - Detecting Deepfake Audio in Vishing Attacks
Detects AI-generated deepfake audio in vishing attacks by extracting spectral features and classifying samples with machine learning models.
Tags
Updated: 2026-09-24Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Preprocess audio samples
- Extract spectral features
- Classify audio samples
- Analyze temporal artifacts
- Inspect spectrograms visually
- Generate forensic reports
Inputs
- Audio samples
- Reference genuine voice corpus
Outputs
- Classification verdict
- Spectrogram visualizations
- Forensic analysis report
Requirements
- Python 3.9+
- librosa, numpy, scikit-learn, scipy
- FFmpeg
- Minimum 3 seconds of audio