BioAuth System

A desktop biometric authentication system built with WPF, OpenCV, and SQL Server.

BioAuth System is a facial recognition and authentication application developed for Windows. The project combines real-time webcam processing, face detection, biometric enrollment, and database-backed user management into a single desktop application.

Users can create an account, enroll their face through the camera, and authenticate using facial recognition.

The system is built around the LBPH (Local Binary Patterns Histograms) face recognition algorithm using OpenCvSharp, with live video capture handled through AForge.NET and the application interface built with WPF on .NET 8.

Overview

The project follows a simple biometric authentication workflow:

User Registration
        │
        ▼
Face Enrollment
        │
        ▼
Multiple Face Samples
        │
        ▼
Face Processing
        │
        ▼
SQL Server Storage
        │
        ▼
Real-Time Face Recognition
        │
        ▼
Authentication Result

Instead of relying only on a traditional username and password, the application adds biometric facial data as part of the authentication process.

Features

Face Enrollment

The system captures multiple facial samples during the registration process.

Each frame is processed to:

Detect the face
Extract the face region
Convert the image to grayscale
Normalize the image size
Prepare the sample for recognition

Multiple samples help provide variation in facial position and improve the recognition process.

Real-Time Face Recognition

The application processes webcam frames in real time.

Detected faces are compared against the biometric data stored in the database using an LBPH Face Recognizer.

The recognition result is evaluated using a confidence or distance threshold.

Lower Distance → Better Match
Higher Distance → Weaker Match

A configurable threshold determines whether authentication is accepted or rejected.

Face Detection with Haar Cascade

Face detection is handled using OpenCV's Haar Cascade classifier.

The system can use parameters such as:

scaleFactor: 1.1
minNeighbors: 5
minSize: new Size(60, 60)

These values help balance detection speed and accuracy.

WPF Desktop Interface

The application uses WPF to provide a modern Windows desktop interface.

The UI includes:

User registration
Face enrollment
Live camera preview
Facial authentication
Application navigation
Animated interface elements

SQL Server Integration

User information and facial image data are stored in Microsoft SQL Server.

The biometric images are stored as binary data using:

VARBINARY(MAX)

This allows the application to manage biometric samples directly through the database rather than depending on external image file paths.

Resource Management

The application handles camera streams and OpenCV resources carefully to reduce unnecessary memory usage during continuous video processing.

Tech Stack

Technology	Purpose
.NET 8	Application framework
WPF	Desktop user interface
C#	Core application logic
OpenCvSharp4	Computer vision and image processing
OpenCvSharp4.Extensions	OpenCV and .NET image integration
LBPHFaceRecognizer	Facial recognition
Haar Cascade	Face detection
AForge.Video.DirectShow	Webcam access and video streaming
Microsoft SQL Server	User and biometric data storage
XAML	Interface design and animations

System Architecture

The application consists of several connected layers:

┌─────────────────────────────┐
│        WPF Interface        │
│ Registration / Login / UI   │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│     Camera Processing       │
│  AForge + OpenCV Pipeline   │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│       Face Detection        │
│        Haar Cascade         │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│      Face Recognition       │
│     LBPHFaceRecognizer      │
└──────────────┬──────────────┘
               │
               ▼
┌─────────────────────────────┐
│         SQL Server          │
│ Users + Biometric Samples   │
└─────────────────────────────┘

Database Structure

The system uses two main entities:

Users

Stores the primary user information.

CREATE TABLE Users (
    UserID INT IDENTITY(1,1) PRIMARY KEY,
    FullName NVARCHAR(100) NOT NULL,
    Username NVARCHAR(50) UNIQUE NOT NULL,
    Password NVARCHAR(255) NOT NULL,
    IsActive BIT DEFAULT 1,
    CreatedAt DATETIME DEFAULT GETDATE()
);

User Faces

Stores the biometric face samples associated with each user.

CREATE TABLE Details (
    DetailID INT IDENTITY(1,1) PRIMARY KEY,
    UserID INT NOT NULL,
    ImagePath VARBINARY(MAX) NOT NULL,
    CreatedAt DATETIME DEFAULT GETDATE(),

    FOREIGN KEY (UserID)
        REFERENCES Users(UserID)
        ON DELETE CASCADE
);

The relationship can be represented as:

Users
  │
  │ 1
  │
  └───────────< Many
                   │
                   ▼
                Details
            Face Samples

A single user can have multiple facial samples stored for training and recognition.

Face Processing Pipeline

During face enrollment, the camera stream passes through the following stages:

Webcam Frame
      │
      ▼
Face Detection
      │
      ▼
Face Region Extraction
      │
      ▼
Grayscale Conversion
      │
      ▼
Histogram Equalization
      │
      ▼
Image Resize
      │
      ▼
100 × 100 Face Sample
      │
      ▼
Database Storage

The processed face samples can be normalized to:

100 × 100 pixels

Histogram equalization can also be applied using:

Cv2.EqualizeHist(source, destination);

This helps normalize image contrast before recognition.

Authentication Workflow

The authentication process works as follows:

Camera Input
     │
     ▼
Detect Face
     │
     ▼
Crop Face Region
     │
     ▼
Convert to Grayscale
     │
     ▼
Normalize Image
     │
     ▼
LBPH Recognition
     │
     ▼
Confidence Evaluation
     │
     ├── Match Accepted
     │
     └── Match Rejected

The recognition threshold can be configured depending on the quality of the dataset and camera conditions.

An example threshold might be:

const double RecognitionThreshold = 80;

In general:

Distance < Threshold  → Authentication Accepted
Distance ≥ Threshold  → Authentication Rejected

The exact threshold should be tested and calibrated using real enrollment data.

Getting Started

Requirements

Before running the project, make sure the following are installed:

Windows 10 or Windows 11
Visual Studio 2022
.NET Desktop Development workload
.NET 8 SDK
Microsoft SQL Server or SQL Server Express
SQL Server Management Studio
A webcam

Installation

Clone the Repository

HTTPS

git clone https://github.com/enzo963/BiometricAuthenticationSystem.git
cd BiometricAuthenticationSystem

SSH

git clone git@github.com:enzo963/BiometricAuthenticationSystem.git

GitHub CLI

gh repo clone enzo963/BiometricAuthenticationSystem

Database Setup

Open SQL Server Management Studio.
Create a new database:

CREATE DATABASE BioAuthDB;

Select the database:

USE BioAuthDB;

Create the Users and Details tables using the schema above.
Configure the connection string according to your SQL Server instance.

Example:

Data Source=YOUR_SERVER;
Initial Catalog=BioAuthDB;
Integrated Security=True;
TrustServerCertificate=True;

For SQL Server Express:

YOUR-PC\SQLEXPRESS

Face Detection Resource

The project requires the Haar Cascade XML file:

haarcascade_frontalface_default.xml

Place it inside:

Resources/
└── haarcascade_frontalface_default.xml

Make sure the file is copied to the output directory when building the application.

Running the Application

Open the solution file in Visual Studio:

BiometricAuthenticationSystem.sln

Then:

Restore the NuGet packages.
Verify the SQL Server connection.
Confirm that the Haar Cascade file is available.
Connect a webcam.
Build the project.
Press F5.

Project Structure

The structure may vary depending on the current implementation, but the project is generally organized around the following components:

BiometricAuthenticationSystem/
│
├── Resources/
│   └── haarcascade_frontalface_default.xml
│
├── Windows/
│   ├── RegistrationWindow.xaml
│   ├── SaveYourFace.xaml
│   ├── LoginWindow.xaml
│   ├── LoginFaceWindow.xaml
│   └── DashboardWindow.xaml
│
├── Services/
│   ├── Database/
│   ├── FaceDetection/
│   └── FaceRecognition/
│
├── App.xaml
├── App.xaml.cs
└── BiometricAuthenticationSystem.sln

Configuration Notes

Database Naming Consistency

The project should use consistent naming for the biometric table and image column.

For example:

Table: Details
Column: ImagePath

or:

Table: UserFaces
Column: FaceData

Both naming styles should not be mixed across different application components.

A single schema should be used throughout the project.

Centralized Connection String

Instead of repeating the connection string in multiple classes, it is recommended to centralize the database configuration.

For example:

App.config

or:

DbConfig.cs

This makes the application easier to configure when the SQL Server instance changes.

Recognition Parameters

The LBPH recognizer can be configured with parameters such as:

Radius
Neighbors
Grid X
Grid Y
Threshold

A common starting configuration is:

Radius: 1
Neighbors: 8
Threshold: 80

These values should be treated as starting points rather than universal values.

Recognition performance depends on:

Camera quality
Lighting conditions
Number of training images
Face angle
Facial expression
Image preprocessing
Dataset quality

Security Considerations

This project is designed as an educational biometric authentication system.

For a production-grade authentication system, additional security measures should be considered:

Password hashing using a modern password hashing algorithm
Encryption for biometric data at rest
Secure database credentials
Rate limiting for authentication attempts
Liveness detection
Anti-spoofing mechanisms
Improved biometric recognition models
Audit logging

LBPH is a classical computer vision algorithm and is useful for learning and experimentation, but it should not be considered sufficient on its own for high-security biometric systems.

Future Improvements

Potential improvements include:

Deep learning-based face recognition
Face embedding models
Liveness detection
Anti-spoofing detection
Camera quality validation
Multi-user recognition
Encrypted biometric templates
Role-based authentication
Authentication logs
REST API integration
Centralized database configuration
Improved application architecture using MVVM
Unit testing
Docker-based database setup

License

This project is licensed under the MIT License.

The project also uses OpenCV resources and third-party libraries distributed under their respective licenses.

Authors

ENZO Abod

GitHub: @enzo963

Disclaimer

BioAuth System was developed as an educational and experimental project for exploring biometric authentication, computer vision, and desktop application development.

Name		Name	Last commit message	Last commit date
Figam		Figam
Resources		Resources
.gitattributes		.gitattributes
.gitignore		.gitignore
App.xaml		App.xaml
App.xaml.cs		App.xaml.cs
AssemblyInfo.cs		AssemblyInfo.cs
Bio Athun System.csproj		Bio Athun System.csproj
Bio Athun System.sln		Bio Athun System.sln
DashboardWindow.xaml		DashboardWindow.xaml
DashboardWindow.xaml.cs		DashboardWindow.xaml.cs
LoginWindow.xaml		LoginWindow.xaml
LoginWindow.xaml.cs		LoginWindow.xaml.cs
README.md		README.md
RegistrationWindow.xaml		RegistrationWindow.xaml
RegistrationWindow.xaml.cs		RegistrationWindow.xaml.cs
SaveYourFace.xaml		SaveYourFace.xaml
SaveYourFace.xaml.cs		SaveYourFace.xaml.cs
loginFaceWindow.xaml		loginFaceWindow.xaml
loginFaceWindow.xaml.cs		loginFaceWindow.xaml.cs
ٍSign_up_page.xaml		ٍSign_up_page.xaml
ٍSign_up_page.xaml.cs		ٍSign_up_page.xaml.cs

Latest commit

History

Folders and files

Repository files navigation