VoiceAuth is an audio deepfake detection application designed to identify manipulated audio content. Utilizing advanced machine learning models, the application processes audio files and provides insights into their authenticity. It supports various audio and video formats, converts them to WAV, and extracts features for analysis. The application has a built-in database and works on Windows.
Features
Deepfake Detection: Uses both a Random Forest model and a Hugging Face pipeline model for accurate predictions.
File Format Support: Handles multiple audio formats (e.g., MP3, WAV, FLAC) and video formats (e.g., MP4, AVI) by converting them to WAV.
MFCC Visualization: Visualizes Mel-Frequency Cepstral Coefficients (MFCC) features extracted from audio files.
Metadata Storage: Logs file metadata, including format, size, audio length, and prediction results in a SQLite database.
User-Friendly Interface: Built with customtkinter, providing a modern and intuitive user experience.
Batch Processing: Allows users to upload and process multiple files simultaneously.
Logging with Typewriter Effect: Displays logs with a typewriter effect for better readability.
VoiceAuth is an audio deepfake detection application designed to identify manipulated audio content. Utilizing advanced machine learning models, the application processes audio files and provides insights into their authenticity. It supports various audio and video formats, converts them to WAV, and extracts features for analysis. The application has a built-in database and works on Windows.
Features
https://github.com/sadiqkassamali/VoiceAuth