This repository contains the official presentation slides for the paper:
Revolutionizing Traffic Management with AI-Powered Machine Vision: A Step Toward Smart Cities
Presented at:
The First Biennial National Conference on the Application of Artificial Intelligence in Traffic Control
📍 University of Isfahan, Iran
📅 6–8 Esfand 1403 (Feb. 25–26, 2025)
- Language: Persian (Farsi)
- Format: PDF (converted from PowerPoint)
- Presenter: Seyed Hossein Hosseini DolatAbadi
- Duration: ~20 minutes
- Audience: Academic & Research Community (AI, Traffic Control, Smart Cities)
The presentation introduces an AI-driven traffic management system based on machine vision and deep learning, focusing on:
- Urban traffic challenges
- Vehicle detection and counting
- Traffic anomaly recognition
- Congestion analysis
- Real-time driver notification
- Smart city infrastructure integration
The system is designed to improve traffic flow, road safety, and driver awareness using real-world surveillance data.
| Slide | Topic |
|---|---|
| 1 | Title & Authors |
| 2 | Introduction: Urban Traffic Challenges |
| 3 | Research Objectives |
| 4 | Methodology Overview |
| 5–6 | Data Collection |
| 7–8 | Data Labeling |
| 9–10 | Preprocessing & Data Augmentation |
| 11–12 | Model Selection & Training (YOLOv8 / YOLOv11) |
| 13 | Backend Analysis & Driver Notification |
| 14–16 | Evaluation & Performance Comparison |
| 17 | Conclusion & Achievements |
| 18 | Accessibility & Participation |
| 19 | References |
| 20 | Q&A |
- Models: YOLOv8, YOLOv11
- Input Size: 640×640
- Training Epochs: 100
- Optimizer: AdamW
- Acceleration: Automatic Mixed Precision (AMP)
- Evaluation Metrics: Precision, Recall, mAP
| Model | Precision | Recall | mAP@50 | mAP@50–95 |
|---|---|---|---|---|
| YOLOv8 | 86.1% | 73.0% | 87.4% | 68.2% |
| YOLOv11 | 89.7% | 72.3% | 81.3% | 62.4% |
YOLOv8 demonstrated better overall performance and adaptability, making it more suitable for real-world traffic deployment.
.
├── Presentation.pdf # Final conference presentation slides
└── README.md
-
Seyed Hossein Hosseini DolatAbadi – University of Isfahan
📧 s.h.hosseini@mehr.ui.ac.ir -
Sayyed Mohammad Hossein Hashemi – University of Isfahan
📧 mhtrxz@gmail.com -
Mohammad Hosseini – Ilam University
📧 40013119812@ilam.ac.ir -
Moein-Aldin AliHosseini – University of Isfahan
📧 moeinaldin2022@gmail.com
- 📄 Paper & Manuscript Repository
- 💻 Source Code & Experiments
- 📦 Dataset Repository
- 🎥 Demo & Qualitative Results
All repositories are maintained under the same GitHub Organization.
This presentation is provided for academic and research purposes only.
For reuse, redistribution, or commercial usage, please contact the authors.
This presentation highlights a practical step toward deploying AI-powered machine vision systems in intelligent transportation infrastructures, contributing to safer, smarter, and more efficient urban environments.