Holistic Recognition of Low Quality License Plates by CNN using Track Annotated Data [IWT4S-AVSS 2017]

Jakub Špaňhel, Jakub Sochor, Roman Juránek, Adam Herout, Lukáš Maršík, Pavel Zemčík
GRAPH@FIT, Brno University of Technology
Corresponding authors: {ispanhel,herout} [at] fit.vutbr.cz

Abstract

This work is focused on recognition of license plates in low resolution and low quality images. We present a methodology for collection of real world (non-synthetic) dataset of low quality license plate images with ground truth transcriptions. Our approach to the license plate recognition is based on a Convolutional Neural Network which holistically processes the whole image, avoiding segmentation of the license plate characters. Evaluation results on multiple datasets show that our method significantly outperforms other free and commercial solutions to license plate recognition on the low quality data. To enable further research of low quality license plate recognition, we make the datasets publicly available.

Downloads

  • Link to paper (TBD)
  • ReId dataset, HDR dataset:
    • The datasets are intended for academic research and non-commercial use only.
    • If you use the dataset, please cite our work.
    • Send an email to ispanhel [at] fit.vutbr.cz for dataset link.

Datasets License


Except where otherwise noted, this work is licensed under
https://creativecommons.org/licenses/by-nc-sa/4.0/legalcode.
© 2017, ŠPAŇHEL, SOCHOR, JURÁNEK, HEROUT, MARŠÍK, ZEMČÍK. Some Rights Reserved.

Citation

@InProceedings{Spanhel2017, 
 author={\v{S}pa\v{n}hel J. and Sochor J. and Jur\'{a}nek R. and Herout A. and Mar\v{s}\'{i}k L. and Zem\v{c}\'{i}k P.}, 
 booktitle={2017 14th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS)}, 
 title={Holistic Recognition of Low Quality License Plates by CNN using Track Annotated Data}, 
 year={2017}, 
 month={Aug}
}

 

Acknowledgements

This work was supported by The Ministry of Education, Youth and Sports of the Czech Republic from the National Programme of Sustainability (NPU II); project IT4Innovations excellence in science – LQ1602. Also, this work was supported by TACR project “RODOS”, TE01020155.