README.md

metadata

language: en
license: mit
tags:
  - yolov8
  - object-detection
  - license-plate
  - arabic-text
  - tunisia
  - computer-vision
library_name: ultralytics
pipeline_tag: object-detection
datasets:
  - Safe-Drive-TN/tunis-word-tunisian-license-plate

Tunisian License Plate - Arabic Text Detection (YOLOv8s)

This model detects the Arabic word "تونس" (Tunis) in Tunisian license plates using YOLOv8s.

Model Description

• Model Type: YOLOv8s (Small)
• Task: Object Detection
• Classes: 1 class - "tunis" (Arabic text region)
• Purpose: Detecting and localizing the word "تونس" in Tunisian license plates for OCR preprocessing

Use Case

This model is designed to be used as a preprocessing step for license plate OCR:

1. Detect the Arabic text "تونس" region
2. Mask or crop this region
3. Apply OCR on the remaining numeric characters for better accuracy

Training Details

• Base Model: YOLOv8s pretrained weights
• Image Size: 512x512
• Framework: Ultralytics YOLOv8
• Training Dataset: Tunisian license plate images

Usage

from ultralytics import YOLO

# Load the model
model = YOLO("yassine-mhirsi/tunis-word-detection-yolov8s")

# Run inference
results = model.predict("path/to/license_plate.jpg", conf=0.5)

# Process results
for result in results:
    boxes = result.boxes
    for box in boxes:
        print(f"Confidence: {box.conf[0]:.2f}")
        print(f"Bounding Box: {box.xyxy[0]}")

Model Files

• best.pt - Best weights from training
• last.pt - Last checkpoint
• Training metrics and visualizations included

Example

Citation

If you use this model, please cite:

@misc{tunis-word-detection-yolov8s,
  authors = {Yassine Mhirsi,Malek Messaoudi},
  title = {Tunisian License Plate Arabic Text Detection},
  year = {2025},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/Safe-Drive-TN/tunis-word-detection-yolov8s/}}
}

License

MIT License