Skip to content

Voice Recognition System using MFCC features and LSTM models for individual identification

Notifications You must be signed in to change notification settings

yeftakun/voice-recognition

Repository files navigation

Voice Recognition

Dataset

Rekam dataset

(mic saya burik, jadi rada" pas nyoba):

  1. file enrol.py, masukan dulu nama individunya.
  2. Setelah tekan "Mulai Perekaman" suara langsung direkam. Tekan "Stop" untuk berhenti.
  3. Ulangi langkah 2 untuk menambahkan file suara.
  4. Upload gambar profile.
  5. Simpan data untuk menyimpan individu. Batal untuk menghapus.

Buat dataset:

  1. Audio file yang mengandung voice karakter/individunya
  2. Remove noise/bgm di Vocal Remover
  3. Tempatkan dataset seperti dibawah
dataset/
│
├── bocchi/
│   ├── voice1.mp3
│   ├── voice2.mp3
│   ├── voice3.mp3
│   └── profile.jpg
│
├── nijika/
│   ├── voice1.mp3
│   ├── voice2.mp3
│   ├── voice3.mp3
│   └── profile.jpg
│
└── susman-from-mic/
    ├── voice1.mp3
    ├── voice2.mp3
    ├── voice3.mp3
    ├── voice4.mp3
    └── profile.jpg

Memulai

  1. Sesuaikan file .env.
  2. Jalankan training.py.
  3. Pastikan sudah menambahkan file voice input di audio/ dan sesuaikan INPUT_FILE pada .env
  4. Jalankan main.py untuk testing.

About

Voice Recognition System using MFCC features and LSTM models for individual identification

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages