Online Exam Monitoring System

Overview

This project implements a full-stack platform for creating, delivering, and proctoring online examinations. It provides separate interfaces for administrators (teachers) and candidates (students), backed by REST APIs and AI-powered monitoring.

Administrators can:

• Register for a subject
• Create exams with multiple-choice questions
• View candidate scores
• Audit exam and cheating logs

Candidates can:

• Register with a photo
• Sign in securely
• Join assigned exams
• Answer questions while a live webcam stream is analysed to detect cheating

The system records suspicious behaviour (e.g., looking away, mobile phone usage) and stores it in a database for later review.

---

System Architecture

Admin Frontend (React + Vite)
        |
        v
Admin Backend (Node.js + Express)
        |
     MongoDB Atlas
        |
Candidate Backend (Python + Flask)
        |
Candidate Frontend (React + Vite)
        |
AI Proctoring (YOLOv5 + Face Recognition + Head Pose)

---

Technology Stack

Domain	Frontend	Backend	Key Libraries & Tools
Admin	React 18 + Vite (JS)	Node.js + Express	Express, Mongoose, JWT, bcrypt, dotenv
Candidate	React + Vite (JS)	Python 3.9 + Flask	Flask, Flask-JWT-Extended, Flask-CORS
Database	—	MongoDB Atlas	Mongoose, PyMongo

---

AI & Proctoring Models

• YOLOv5s (Ultralytics)
  Detects mobile phones and multiple people in webcam frames.

• MediaPipe FaceMesh
  Estimates head pose and flags left/right head movement beyond ±10°.

• Face Recognition (dlib-based)
  Verifies that the registered candidate is the one present during the exam.

All cheating events are logged with timestamps and usernames.

---

Features

Administrator Features

• Secure registration and login (JWT + bcrypt)
• Subject-wise exam creation
• Dynamic MCQ management
• Dashboard for exams, scores, and logs
• Cheating log review

Candidate Features

• Registration with face image
• Secure login
• Live exam interface
• AI-based webcam monitoring
• Automatic cheating detection
• Answer submission and scoring

---

Project Structure

Online-Exam-Monitoring-System/
├── admin/
│   ├── backend/
│   │   ├── models/
│   │   │   └── Admin.js
│   │   ├── routes/
│   │   │   ├── auth.js
│   │   │   ├── exams.js
│   │   │   └── scores.js
│   │   └── server.js
│   └── frontend/
│       ├── src/
│       │   ├── components/
│       │   └── context/
│       └── package.json
├── user/
│   ├── backend/
│   │   ├── app.py
│   │   ├── onlineproctor.py
│   │   └── requirements.txt
│   └── frontend/
│       └── package.json
└── README.md

---

Prerequisites

• Node.js (v18+)
• Python 3.9
• MongoDB Atlas or Local MongoDB
• Visual Studio Community 2022 (Desktop Development with C++)
• Webcam

---

Setup & Installation

Clone Repository

git clone https://github.com/chauhanavi21/Online-Exam-Monitoring-System.git
cd Online-Exam-Monitoring-System

Admin Frontend

cd admin/frontend
npm install
npm run dev

Admin Backend

cd admin/backend
npm install
npm run dev

Runs on: http://localhost:5001

Candidate Frontend

cd user/frontend
npm install
npm run dev

Candidate Backend (AI Proctoring)

cd user/backend
python -m venv Exam
# Windows
.\Exam\Scripts\Activate.ps1
# macOS / Linux
source Exam/bin/activate

pip install --upgrade pip
pip install cmake
pip install -r requirements.txt
python -m flask run

---

Running the Application

You need four terminals:

Admin

cd admin/backend
npm run dev

cd admin/frontend
npm run dev

Candidate

cd user/backend
python -m flask run

cd user/frontend
npm run dev

---

Notes

• All data is stored in MongoDB
• Admin backend uses Mongoose
• Candidate backend uses PyMongo
• HTTPS and environment variables are recommended for production

---

Author

Avi Chauhan
GitHub: https://github.com/chauhanavi21