License Plate Detection + Recognition + Identification + Verification

We are going to use deep-text-recognition-benchmark repository to recognize persian license plate numbers .

How to install

pip install -r requirements.txt

Download pre-trained weights

YOLOv8 detector

DTRB recognizer

How to use

python main.py --recognizer_weights weights\dtrb_recognizer\Dtrb_TPS-ResNet-BiLSTM-Attn_License_Plate_Recognizer.pth --input_image io\input\image3.jpg --threshold 0.5

This command will run the main.py file and pass the values for the --recognizer_weights, --input_image, and --threshold arguments as specified in your code. Make sure to navigate to the directory where your main.py file is located before running this command.

---

Results of license plate verification

Ground Truth	predicted_labels	confidence score
	76t16522	0.9975
	38e47755	0.9854

---

Ground Truth	predicted_labels	confidence score
	62q97366	0.9986
	3m1970	0.1528

---

Ground Truth	predicted_labels	confidence score
	89q56966	0.9948
	34b29855	0.7967