Speaker-Verification-GUI

🎤 Speaker Verification GUI (SpeechBrain ECAPA-TDNN)

An intuitive and efficient GUI-based system for real-time speaker registration and verification using the SpeechBrain library and ECAPA-TDNN model.

---

🚀 Features

✅ Register speaker from audio file
✅ Record voice and register directly
✅ Identify speaker from file or real-time recording
✅ Auto-register all files in ref_voices/
✅ View logs, delete speakers, and adjust threshold
✅ Modern Tkinter interface (VSCode-inspired style)

GUI_3

GUI_2

GUI_1

---

🧠 Model Information

• Model: speechbrain/spkrec-ecapa-voxceleb
• Downloaded automatically from Hugging Face on first run
• Stored locally in pretrained_models/ folder for reuse (no need to redownload)

---

🛠️ Installation

git clone https://github.com/Mrkomiljon/Speaker-Verification-GUI.git
cd Speaker-Verification-GUI

Install dependencies

pip install -r requirements.txt

▶️ Usage

python speaker_verification_gui_1/2/3.py

What You Can Do: 🎙 Record your voice → auto-save to ref_voices/ → register

📁 Select any .wav or .mp3 → identify speaker

🧹 Clear logs, ❌ Delete speakers, 🔧 Adjust threshold

📌 Notes

Ensure your microphone is enabled and default device is selected.

The minimum required audio duration is 2 seconds for accurate embedding.

Supported formats: .wav, .mp3

💡 Future Plans

Multi-model selection (e.g., ECAPA vs. Titanet)

Speaker statistics and timeline

CRM integration (e.g., saving call logs or speaker names)

📄 License

This project is open-source and available under the MIT License.

✨ Acknowledgements

SpeechBrain

ECAPA-TDNN Paper

Tkinter + ttk for UI