Detects AI-generated (deepfake) voices using MFCC features with CNN, Random Forest, and KNN models. Achieves ~98% accuracy using engineered audio features and the SceneFake dataset.
random-forest voice cnn knn-classification spectogram deepfake-detection mfcc-features audio-deepfake-detection melscalogram cnn-audio-recognition
-
Updated
Nov 29, 2025 - Jupyter Notebook