In EDS ansehen

Weighted Edit Distance for Country Code Recognition in License Plates

Gespeichert in:
Bibliographische Detailangaben
Titel: Weighted Edit Distance for Country Code Recognition in License Plates
Autoren: Kateryna Chumachenko, Alexandros Iosifidis, Moncef Gabbouj
Verlagsinformationen: Zenodo
Publikationsjahr: 2022
Bestand: Zenodo
Schlagwörter: license plate recognition, edit distance, Leven- shtein distance, country code recognition
Beschreibung: This paper presents the problem of country code recognition from license plate images. We propose an approach based on character detection and subsequent clustering for country code localization. We further propose three weighted Edit Distance metrics for country of origin prediction from imperfect detections, namely based on character similarity, detection confidence, and relative operation importance. Experimental results show the benefit of proposed approaches on real-world data. The proposed method is lightweight and independent of the underlying object detector, facilitating its application on edge devices. ; This work was supported by Business Finland project 5G-VIIMA. A. Iosifidis acknowledges funding from the EU H2020 research and innovation program under grant agreement No. 957337 (MARVEL).
Publikationsart: conference object
Sprache: English
Relation: https://zenodo.org/communities/marvel_project/; https://zenodo.org/communities/eu/; https://zenodo.org/records/6736842; oai:zenodo.org:6736842; https://doi.org/10.5281/zenodo.6736842
DOI: 10.5281/zenodo.6736842
Verfügbarkeit: https://doi.org/10.5281/zenodo.6736842
https://zenodo.org/records/6736842
Rights: Creative Commons Attribution 4.0 International ; cc-by-4.0 ; https://creativecommons.org/licenses/by/4.0/legalcode
Dokumentencode: edsbas.E8DBD50B
Datenbank: BASE
Beschreibung
Abstract:This paper presents the problem of country code recognition from license plate images. We propose an approach based on character detection and subsequent clustering for country code localization. We further propose three weighted Edit Distance metrics for country of origin prediction from imperfect detections, namely based on character similarity, detection confidence, and relative operation importance. Experimental results show the benefit of proposed approaches on real-world data. The proposed method is lightweight and independent of the underlying object detector, facilitating its application on edge devices. ; This work was supported by Business Finland project 5G-VIIMA. A. Iosifidis acknowledges funding from the EU H2020 research and innovation program under grant agreement No. 957337 (MARVEL).
DOI:10.5281/zenodo.6736842