Skip to content

flask website with GMM working as anomaly detection model to predict who is taking and if he is saying the right sentence or not.

Notifications You must be signed in to change notification settings

youssef-shaban/voice-fingerprint-lock

 
 

Repository files navigation

Voice-Passsword-Checker || DSP Task 3

Table of contents:

Introduction

Voice Password checker is a web application used for Voice authentication (voice biometrics voice ID or speaker recognition). The system is speech dependent which requires authorized person with correct sentence said. In our appplication, Team Members are only authorized with "Open the door" sentence.

Work Process

  • Voice Feature Extraction depending on MFCCs:

  • Models Training for user and sentence identification

  • Model Deployment using web technologies (Flask, Html, CSS, JS)

Project Features

In this web application you can

  1. Easily use recorder to check
  1. Live Visulaization for input audio features
  1. Models Scores of input audio (User and Sentence)

Project full Demo

Project Structure

The Web Application is built using:

  • Frontend:
    • HTML
    • CSS
    • JavaScript
    • Ajax
  • Backend framework:
    • Flask (Python)

The Frontend main function to set the structure of the page and plot the signals and mange the user interface while the backend function is to do on-signal operations like sampling, reconstructing, adding signal, adding noise & removing signal.

main
├─ Notebooks
├─Training sets
├─ static (JS & CSS files)
│  ├─  css
│  ├─  img
│  └─  js
├─ template (HTML files)
├─ app.py (Back-End Server)
├─ Gmm.py (Audio preprocessing and Models Predictions)
├─ utilities.py (Data to draw helpers)
└─ README.md

Run the Project

  1. Install Python3 in your computer
Download it from www.python.org/downloads/
  1. Install the following packages
    • numpy
    • flask
    • flask_cors
    • Librosa
    • pickle
    • scipy
    • sklearn
    • python_speech_features
  • Open Project Terminal & Run
pip install -r requirments.txt
  1. Start Server by Running
python app.py
  1. Visit http://127.0.0.1:5001

Team

First Semester - Biomedical Digital Signal Processing (SBE3110) class project created by:

Team Members' Names Section B.N.
Dina Hussam 1 28
Romaisaa Shrief 1 36
Youssef Shaaban 2 56

Submitted to:

  • Dr. Tamer Basha & Eng. Abdullah Darwish All rights reserved © 2022 to Team 14 - Systems & Biomedical Engineering, Cairo University (Class 2024)

About

flask website with GMM working as anomaly detection model to predict who is taking and if he is saying the right sentence or not.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Jupyter Notebook 97.4%
  • Other 2.6%