This repository provides the materials to experiment with the CityDB dataset for License Plate Matching (LPM)

Dataset Download Link: https://drive.google.com/file/d/1WJ2PmAKpCBLFE1rk11k_ZT41ga7-Lm1G/view?usp=share_link

The Notebook LPM_Evaluate_CityDB.ipynb demonstrate how to evaluate the code for obtaining results on this dataset. We have provided pre-computed CNN features in the file CityDB_features_VGGS.npz. You can replace it with your own features and generate results using this notebook.

Please use the following reference if you exploit the CityDB dataset in your research. Moreover, you can cite this if you would like to use the evaluation procedures mentioned in this paper.

Hasnat, A., and Nakib, A. "Robust license plate signatures matching based on multi-task learning approach." Neurocomputing 440 (2021): 58-71.

@article{hasnat2021robust, title={Robust license plate signatures matching based on multi-task learning approach}, author={Hasnat, Abul and Nakib, Amir}, journal={Neurocomputing}, volume={440}, pages={58--71}, year={2021}, publisher={Elsevier} }