synthetic-speech-detection 开源项目推荐（12 个）

ductuantruong /tcm_add

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

elleros /spoofed-speech-detection

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

Ashigarg123 /ShiftySpeech

该项目在 GitHub 上没有填写描述

changtaoli /oct-for-spoofing-detection

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

pujariaditya /WaveVerify

Audio watermarking system for deepfake detection and media authentication.

umitkacar /ensemble-speaker-verification

合奏扬声器验证达到97%准确率——MFCC+DTW（92%）和Resemblyzer CNN（94%）的智能融合用于语音认证

wobushannes /voice-deepfake-detector-gui

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

Security-FIT /SpAArSIST

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

blindconf /fingerprint

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

venkatasriramt-spec /AI-Voice-Forensics

Real-time multilingual AI voice deepfake & voice cloning detector across 9 languages. Fuses WavLM embeddings, LFCC spectral forensics, and LightGBM with sub-15ms inference (~99.8% in-distribution benchmark). Includes REST API & WebSockets.

Security-FIT /hidden-cost-pairwise-verification

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

matiuste /RSF

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