CertifySecure: Blockchain-Integrated Student Certificate Validation App with Flutter

Flask-Based Face Recognition

🚀 Project Overview

CertifySecure is a Flask-based face recognition system utilizing OpenCV and Haarcascade for secure and efficient student authentication. This backend processes biometric authentication requests and integrates Firebase for data storage and authentication.

🔒 Key Features

Face Recognition Authentication

• Uses OpenCV and Haarcascade for face detection and recognition.
• Low-latency recognition process ensuring quick authentication.
• Biometric security to prevent unauthorized access.
• Machine learning-based feature extraction for accurate identification.

Firebase Integration

• Stores student authentication data securely.
• Firebase Admin SDK for user management and authentication.
• Real-time authentication logs for security monitoring.

🨠 Technical Architecture

Updated Project Structure

C:\face-recognition-app\app\
├── main.py              # Face recognition processing
├── setup_local.py       # Local setup script
├── train.py             # Model training script
├── utils/               # Utility functions
├── dataset/             # Training dataset storage
├── model/               # Trained models storage
├── secrets/
│   └── firebase-credentials  # Firebase credentials stored securely (no .json extension)
├── temp/                # Temporary storage for image processing
└── .env                 # Environment variables configuration

Technology Stack

• Backend: Flask (Python)
• Face Recognition: OpenCV, Haarcascade, Scikit-learn
• Authentication & Storage: Firebase, Google Drive
• Environment Management: Python-dotenv

📸 Face Recognition Methodology

Detection Techniques

• Haar Cascade Classifiers (for face detection)
• OpenCV-based feature extraction
• Scikit-learn for encoding and comparison

Recognition Process

1. Capture Image Input (Sent via API request)
2. Face Detection (Using OpenCV Haarcascade)
3. Feature Extraction & Encoding
4. Compare Encoded Features with Stored Data
5. Return Authentication Status

🚀 Installation & Setup

Prerequisites

• Python 3.9+
• Firebase Project with Admin SDK Credentials
• Google Drive API Access

Backend Setup

# Clone Repository
git clone https://github.com/your-org/certifysecure-face-recognition.git
cd certifysecure-face-recognition

# Create Virtual Environment
python -m venv venv
source venv/bin/activate  # On Windows use `venv\Scripts\activate`

# Install Dependencies
pip install -r requirements.txt

# Set Up Environment Variables
cp .env.example .env  # Configure Firebase and Google Drive credentials

# Run Flask App
python main.py

🔐 Security Measures

• Biometric Authentication (Face Recognition)
• Firebase Authentication for secure access management
• Environment Variable Encryption to protect sensitive keys

🤖 Machine Learning Models

Face Recognition Models

• Haar Cascade Classifiers for face detection
• LBPH (Local Binary Patterns Histograms) for recognition

Model Training

• Dataset Preprocessing: Image resizing, grayscale conversion
• Encoding & Training: Feature extraction & storage in label.pkl
• Retraining for Improved Accuracy

📚 Dataset Training Process

Dataset Creation and Training

The dataset training process involves capturing student facial images, preprocessing them, and training a Local Binary Pattern Histogram (LBPH) model for recognition.

🔹 Step 1: Capture Training Data

import cv2
import os
import numpy as np
from datetime import datetime

class DatasetCreator:
    def __init__(self):
        self.face_cascade = cv2.CascadeClassifier(
            cv2.data.haarcascades + "haarcascade_frontalface_default.xml"
        )
        self.cap = cv2.VideoCapture(0)
    
    def create_user_dataset(self, user_id, num_images=500):
        dataset_path = f"dataset/{user_id}"
        os.makedirs(dataset_path, exist_ok=True)
        count = 0
        while count < num_images:
            ret, frame = self.cap.read()
            gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
            faces = self.face_cascade.detectMultiScale(gray, scaleFactor=1.3, minNeighbors=5, minSize=(30, 30))
            for (x, y, w, h) in faces:
                face = gray[y:y+h, x:x+w]
                face_resized = cv2.resize(face, (200, 200))
                timestamp = datetime.now().strftime("%Y%m%d_%H%M%S_%f")
                cv2.imwrite(f"{dataset_path}/{timestamp}.jpg", face_resized)
                count += 1

💪 Training Requirements

• Minimum Images: 500 per student
• Conditions: Well-lit environment, various facial expressions
• Recognition Confidence Threshold: <100 (lower is better)

📊 Performance Metrics

Metric	Value
Image Resolution	200x200 px
Training Dataset Size	500+ imgs
Face Detection Accuracy	>90%
Recognition Accuracy	85-95%

🌐 API Endpoints

Face Recognition Endpoint

@app.route('/recognize', methods=['POST'])
def recognize_face():
    # Capture and process face
    # Compare with stored data
    # Verify against Firebase authentication records
    # Return authentication status

🌍 Deployment Options

• Docker Deployment for containerized execution
• Google Cloud Run / AWS Lambda for scalable hosting
• Render for easy deployment with auto-scaling

🌐 Setting Up Google Drive API

1. Enable Google Drive API in Google Cloud Console.
2. Create a Service Account and Generate JSON Key.
3. Store credentials.json securely in the project.

🔒 Obtaining Firebase Service Account JSON File

1. Go to Firebase Console.
2. Select Your Project and navigate to Service Accounts.
3. Generate a new private key and download it.

📚 Future Enhancements

• QR Code Verification for instant certificate validation.
• UI Enhancements for better user experience.
• Face Recognition with Deep Learning Models for improved accuracy.

📞 Contact & Support

• Email: [email protected]
• LinkedIn: Sri Nitish Kamisetti
• GitHub: Nitish2773