Publication detail

Perceptual license plate super-resolution with CTC loss

BÍLKOVÁ, Z. HRADIŠ, M.

Original Title

Perceptual license plate super-resolution with CTC loss

Type

conference paper

Language

English

Original Abstract

We present a novel method for super-resolution (SR) of license plate images based on an end-to-end convolutional neural networks (CNN) combining generative adversial networksn(GANs) and optical character recognition (OCR). License plate SR systems play an important role in number of security applications such as improvement of road safety, traffic monitoring or surveillance. The specific task requires not only realistic-looking reconstructed images but it also needs to preserve the text information. Standard CNN SR and GANs fail to accomplish this requirment. The incorporation of the OCR pipeline into the method also allows training of the network without the need of ground truth high resolution data which enables easy training on real data with all the real image degradations including compression.

Keywords

superresolution, license plate recognition, GAN, deblurring

Authors

BÍLKOVÁ, Z.; HRADIŠ, M.

Released

15. 1. 2020

Publisher

Society for Imaging Science and Technology

Location

Springfield, USA

ISBN

2470-1173

Year of study

2020

Number

6

Pages from

52

Pages to

57

Pages count

5

BibTex

@inproceedings{BUT182964,
  author="Zuzana {Bílková} and Michal {Hradiš}",
  title="Perceptual license plate super-resolution with CTC loss",
  booktitle="IS and T International Symposium on Electronic Imaging Science and Technology",
  year="2020",
  volume="2020",
  number="6",
  pages="52--57",
  publisher="Society for Imaging Science and Technology",
  address="Springfield, USA",
  doi="10.2352/ISSN.2470-1173.2020.6.IRIACV-052",
  issn="2470-1173"
}