Automated_detection_of_texting_while_driving_using_Yolov8_Ultralytics

Texting While Driving Detection Using YOLOv8

A computer vision-based pipeline that detects texting while driving using YOLOv8. Built for real-time monitoring with a focus on privacy compliance (GDPR).

---

📌 Table of Contents

1. Project Overview
2. Dataset
3. Model Architecture
4. Evaluation Metrics
5. Results
6. Installation
7. Usage
8. Future Work
9. Contributors
10. License

---

📖 Project Overview

Texting while driving is a major cause of road accidents. This project presents an automated detection system using the YOLOv8 object detection model to identify drivers who are texting while driving. The system is trained on a curated dataset and shows high precision and recall.

---

📁 Dataset

• Source: Roboflow “Phone While Driving” dataset
• Classes: Phone, Wheel
• Splits:
  • Training: 620 images
  • Validation: 177 images
  • Testing: 89 images

Annotation format: YOLO bounding boxes.

---

🧠 Model Architecture

• Framework: YOLOv8 by Ultralytics
• Components:
  • Backbone: CSPDarknet53 + C2f
  • Neck: FPN + PAN
  • Head: Bounding box and class predictions

---

📊 Evaluation Metrics

Metric	Validation	Test
mAP50-95	0.888	0.855
mAP50	0.987	0.983
Precision	0.967	0.973
Recall	0.981	0.978

---

✅ Results

• High detection accuracy on both Phone and Wheel classes
• Minimal overfitting between validation and test sets
• Clear potential for real-world deployment

---

🛠️ Installation

git clone https://github.com/YOUR_USERNAME/texting-detection-yolov8.git
cd texting-detection-yolov8
pip install -r requirements.txt