Automatic License Plate Blurring System using YOLOv11 β detects license plates in images & videos and applies automated blurring for privacy protection. Includes training, evaluation, and inference pipeline.
# ππ Automatic License Plate Blurring System
Note: π Please read the Documentation for more details.
This project is an **Automatic License Plate Blurring System** built using **YOLO object detection models**.
The system detects license plates in both **images** and **videos**, then applies Gaussian blurring filter for privacy protection.
---
## π Dataset & Format
- **Source:** Egyptian Cars Plates (Kaggle)
- 2,085 labelled images of Egyptian plates
- **Format:** `.csv`
---
## π Features
- π **License Plate Detection** using YOLO
- π **Model Performance Comparison** across different YOLO variants & training epochs
- π Selected the **best performing model** based on **Precision / Recall / mAP** metrics
- π₯ **Real-world Application** on both images and videos
- π Automated **privacy preservation** by blurring detected plates
---
## π§ Model Training
1. Trained **YOLOv8n**, **YOLOv10n**, and **YOLOv11n** models with varying epochs (**3**, **5**, **10**)
2. Evaluated performance using:
- β
Precision
- β
Recall
- β
mAP (Mean Average Precision)
- β
Val_box_loss, Val_cls_loss, Val_cls_loss,
3. Selected the best-performing model (**YOLOv11n**) and trained it to **50 epochs**
4. Used the final model on different **images & videos**
---
## π Results
---
## πΌοΈ Sample Results
---