Adamu Ali-Gombe
Cross domain evaluation of text detection models.
Ali-Gombe, Adamu; Elyan, Eyad; Moreno-García, Carlos; Jayne, Chrisina
Authors
Professor Eyad Elyan e.elyan@rgu.ac.uk
Professor
Dr Carlos Moreno-Garcia c.moreno-garcia@rgu.ac.uk
Senior Lecturer (A)
Chrisina Jayne
Contributors
Elias Pimenidis
Editor
Plamen Angelov
Editor
Christina Jayne
Editor
Antonios Papaleonidas
Editor
Mehmet Aydin
Editor
Abstract
Text detection is a very common task across a wide range of domains, such as document image analysis, remote identity verification, amongst others. It is also considered an integral component of any text recognition system, where the performance of recognition tasks largely depends on the accuracy of the detection of text components. Various text detection models have been developed in the past decade. However, localizing text characters is still considered as one of the most challenging computer vision tasks within the text recognition task. Typical challenges include illumination, font types and sizes, languages, and many others. Furthermore, detection models are often evaluated using specific datasets without much work on cross-datasets and domain evaluation. In this paper, we present an experimental framework to evaluate the generalization capability of state-of-the-art text detection models across different application domains. Extensive experiments were carried using different established methods: EAST, CRAFT, Tessaract and Ensembles applied to various publicly available datasets. The generalisation performance of the models was evaluated and compared using precision, recall and F1-score. This paper opens a future direction in investigating ensemble models for text detection to improve generalisation.
Citation
ALI-GOMBE, A., ELYAN, E., MORENO-GARCÍA, C. and JAYNE, C. 2022. Cross domain evaluation of text detection models. In Pimenidis, E., Angelov, P., Jayne, C., Papaleonidas, A. and Aydin, M. (eds.) Artificial neural networks and machine learning - ICANN 2022: proceedings of the 31st International conference on artificial neural networks (ICANN 2022), 6-9 September 2022, Bristol, UK, part III. Lecture notes in computer science, 13531. Cham: Springer [online], pages 50-61. Available from: https://doi.org/10.1007/978-3-031-15934-3_5
Conference Name | 31st International conference on artificial neural networks 2022 (ICANN22) |
---|---|
Conference Location | Bristol, UK |
Start Date | Sep 6, 2022 |
End Date | Sep 9, 2022 |
Acceptance Date | Jun 20, 2022 |
Online Publication Date | Sep 9, 2022 |
Publication Date | Dec 31, 2022 |
Deposit Date | Sep 13, 2022 |
Publicly Available Date | Sep 10, 2023 |
Publisher | Springer |
Pages | 50-61 |
Series Title | Lecture notes in computer science |
Series Number | 13531 |
Series ISSN | 0302-9743; 1611-3349 |
Book Title | Artificial neural networks and machine learning - ICANN 2022 |
ISBN | 9783031159336 |
DOI | https://doi.org/10.1007/978-3-031-15934-3_5 |
Keywords | Text detection; Efficient and accurate scene text detector; Character aware region awareness for text detection; Tesseract; Ensembles |
Public URL | https://rgu-repository.worktribe.com/output/1752655 |
Files
This file is under embargo until Sep 10, 2023 due to copyright reasons.
Contact publications@rgu.ac.uk to request a copy for personal use.
You might also like
Towards a reliable face recognition system.
(2020)
Conference Proceeding
Multiple fake classes GAN for data augmentation in face image dataset.
(2019)
Conference Proceeding
Few-shot classifier GAN.
(2018)
Conference Proceeding
Face detection with YOLO on edge.
(2021)
Conference Proceeding