Smart Classroom Behavior Detection using YOLOv8

Internship Project Organization:Bharat Electronics Limited (BEL)

Project Overview

This project was developed during my internship at Bharat Electronics Limited (BEL). The main objective of the project is to detect different student behaviors in a classroom using the YOLOv8 object detection model. The system analyzes classroom images and videos and identifies various student activities, making it useful for classroom monitoring and behavior analysis.

Project Objectives

• Detect different student behaviors using Computer Vision.
• Train a custom YOLOv8 model on a classroom dataset.
• Test the model on classroom images and videos.
• Analyze the detection results using evaluation metrics.
• Understand the workflow involved in an AI-based object detection project.

Technologies Used

• Python
• YOLOv8
• OpenCV
• Google Colab
• Roboflow
• Kaggle Dataset
• GitHub

Dataset

The project uses a classroom behavior dataset collected from Kaggle. The dataset was annotated and exported using Roboflow in YOLOv8 format before training the model.

Project Workflow

Dataset Collection ↓ Data Preprocessing ↓ YOLOv8 Model Training ↓ Model Validation ↓ Image & Video Testing ↓ Behavior Detection

Behaviors Detected

The model detects the following classroom activities:

• Hand Raise
• Reading
• Writing
• Sleeping
• Using Device
• Standing
• Looking Forward
• Turning Head

Results

The trained model was able to detect multiple classroom behaviors successfully. The project includes:

• Prediction Images
• Video Detection
• Confusion Matrix
• Precision Curve
• Recall Curve
• Precision-Recall Curve
• Training Results Graph

Repository Structure

📁 documentation
📁 model
📁 notebooks
📁 outputs
📁 screenshots

README.md
requirements.txt

---

Future Improvements

• Improve detection accuracy using a larger dataset.
• Implement real-time webcam detection.
• Develop a classroom analytics dashboard.
• Generate attendance and engagement reports.
• Deploy the project as a web application.

Author

B S Vanisha Babu

B.Tech Computer Science Engineering (AI & ML)

Presidency University, Bengaluru

Project Status: Completed