Voice biometric authentication PAM module for Linux

Voice authentication system implementation using Python

🎙️ Speaker recognition service for Home Assistant using voice embeddings - train on voice samples and identify speakers.

[ICASSP'23] Online speaker clustering

The project is related to the development of labs for the ITMO Speaker Recognition Course.

IDVoice + ChatGPT iOS demo app

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

This repo contains the reicpe to assemble a corpus for Foreign Accented English using the crowdsourced corpus Common Voice which contains (optional) accent labels.

Swift library for Speaker Embedding extraction and verification using NVIDIA NeMo TitaNet model converted to CoreML. Extract 192-dim speaker embeddings, verify speakers, and perform real-time speaker diarization on iOS/macOS.

The FishBoardMix corpus is designed to explore Speaker-Age estimation technology.

🤖 JARVIS-style AI voice assistant with voice biometrics, face recognition, and computer vision. 100% offline using Ollama. Built in Python for macOS.

IDVoice + ChatGPT Android demo app

The project is related to the development of Basics of Voice Biometrics lecture book for the ITMO Speaker Recognition Course.

Lightweight Siamese speaker verification with mel-spectrogram embeddings.

Voice Truthgate™ - 1st Honest, on-device voice-deepfake detection by mosADD™

A collection of scripts and data I used when working on my dissertation

Speech-Driven Secure Banking System with Voice Biometrics using React Native, Flask, Whisper STT and Speaker Verification.

Small speaker recognition demo using SpeechBrain ECAPA-TDNN embeddings, centroid-based cosine scoring, and a FastAPI + browser frontend for file upload and microphone input.

Speaker recognition using Gaussian Mixture Models (GMM) involves identifying speakers by modeling their voice features with statistical distributions. GMMs are trained on speech data to distinguish between different speakers based on their unique vocal characteristics. This technique is commonly used in voice biometrics and authentication systems.