[INTERSPEECH'24] Temporal-Channel Modeling in Multi-head Self-Attention for Synthetic Speech Detection

Implementation of a standard synthetic speech detection approach via bob.bio.spear Python package

Offical implementation of "The Role of Long-term Dependency in Synthetic Speech Detection"

Audio watermarking system for deepfake detection and media authentication.

Ensemble speaker verification achieving 97% accuracy - Intelligent fusion of MFCC+DTW (92%) and Resemblyzer CNN (94%) for voice authentication

Desktop GUI tool for detecting audio deepfakes using hand-crafted features + XGBoost/LightGBM ensemble. Segment-level analysis, heatmaps, PDF reports. Zero deep learning required.

[Paper Under Revision] Lightweight Detection and Model Attribution of Synthetic Speech via Residual Statistical Fingerprints.

Code for "The Hidden Cost of Pairwise Verification in Synthetic Speech Source Tracing"

Sparsified AASIST for Efficient and Reliable Anti-Spoofing — reference implementation (Interspeech 2026).

[Interspeech 2026 Long Track] Lightweight Detection and Model Attribution of Synthetic Speech via Residual Statistical Fingerprints.