Pelee-Text++: A Tiny Neural Network for Scene Text Detection
Scene text detection has become an important field in the computer vision area due to the increasing number of applications. This is a very challenging problem as textual elements are commonly found in “noisy” and complex natural scenes. Another issue refers to the presence of...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2020-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/9289838/ |
id |
doaj-1ece3e7a430b48e6aacdea87eed55b0e |
---|---|
record_format |
Article |
spelling |
doaj-1ece3e7a430b48e6aacdea87eed55b0e2021-03-30T03:42:25ZengIEEEIEEE Access2169-35362020-01-01822317222318810.1109/ACCESS.2020.30438139289838Pelee-Text++: A Tiny Neural Network for Scene Text DetectionManuel Cordova0https://orcid.org/0000-0002-6527-6740Allan Pinto1https://orcid.org/0000-0003-3765-8300Helio Pedrini2https://orcid.org/0000-0003-0125-630XRicardo da Silva Torres3https://orcid.org/0000-0001-9772-263XInstitute of Computing, University of Campinas, Campinas, BrazilInstitute of Computing, University of Campinas, Campinas, BrazilInstitute of Computing, University of Campinas, Campinas, BrazilDepartment of ICT and Natural Sciences, Faculty of Information Technology and Electrical Engineering, NTNU–Norwegian University of Science and Technology, Ålesund, NorwayScene text detection has become an important field in the computer vision area due to the increasing number of applications. This is a very challenging problem as textual elements are commonly found in “noisy” and complex natural scenes. Another issue refers to the presence of texts encoded into different languages within the same image. State-of-the-art solutions rely on the use of deep neural network approaches or even ensembles of them. However, such solutions are associated with “heavy” models, which are computationally expensive in terms of memory and storage footprints, which hampers their use in real-time mobile applications. In this work, we introduce Pelee-Text++, a lightweight neural network architecture for multi-lingual multi-oriented scene text detection, especially tailored to running on devices with computational restrictions. Additionally, to the best of our knowledge, this is the first work to evaluate the performance of text detection methods in commercial smartphones. Over this scenario, Pelee-Text++ processes 2.94 frames per second and it is the only evaluated approach that did not cause memory issues on smartphones, even using an input image of $1024 × 1024 pixels. Our proposal achieves a promising trade-off between efficiency and effectiveness, with a model size of 27 Megabytes and F-measure of 91.20%, 85.78%, 81.72%, 80.30%, 82.53% and 66.51% on ICDAR 2011, ICDAR 2013, ICDAR 2015, MSRA-TD500, ReCTS 2019 and Multi-lingual 2019 datasets, respectively.https://ieeexplore.ieee.org/document/9289838/Text detectionmobile-networkmobile devicesmulti-oriented textmulti-lingualconvolutional neural network |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Manuel Cordova Allan Pinto Helio Pedrini Ricardo da Silva Torres |
spellingShingle |
Manuel Cordova Allan Pinto Helio Pedrini Ricardo da Silva Torres Pelee-Text++: A Tiny Neural Network for Scene Text Detection IEEE Access Text detection mobile-network mobile devices multi-oriented text multi-lingual convolutional neural network |
author_facet |
Manuel Cordova Allan Pinto Helio Pedrini Ricardo da Silva Torres |
author_sort |
Manuel Cordova |
title |
Pelee-Text++: A Tiny Neural Network for Scene Text Detection |
title_short |
Pelee-Text++: A Tiny Neural Network for Scene Text Detection |
title_full |
Pelee-Text++: A Tiny Neural Network for Scene Text Detection |
title_fullStr |
Pelee-Text++: A Tiny Neural Network for Scene Text Detection |
title_full_unstemmed |
Pelee-Text++: A Tiny Neural Network for Scene Text Detection |
title_sort |
pelee-text++: a tiny neural network for scene text detection |
publisher |
IEEE |
series |
IEEE Access |
issn |
2169-3536 |
publishDate |
2020-01-01 |
description |
Scene text detection has become an important field in the computer vision area due to the increasing number of applications. This is a very challenging problem as textual elements are commonly found in “noisy” and complex natural scenes. Another issue refers to the presence of texts encoded into different languages within the same image. State-of-the-art solutions rely on the use of deep neural network approaches or even ensembles of them. However, such solutions are associated with “heavy” models, which are computationally expensive in terms of memory and storage footprints, which hampers their use in real-time mobile applications. In this work, we introduce Pelee-Text++, a lightweight neural network architecture for multi-lingual multi-oriented scene text detection, especially tailored to running on devices with computational restrictions. Additionally, to the best of our knowledge, this is the first work to evaluate the performance of text detection methods in commercial smartphones. Over this scenario, Pelee-Text++ processes 2.94 frames per second and it is the only evaluated approach that did not cause memory issues on smartphones, even using an input image of $1024 × 1024 pixels. Our proposal achieves a promising trade-off between efficiency and effectiveness, with a model size of 27 Megabytes and F-measure of 91.20%, 85.78%, 81.72%, 80.30%, 82.53% and 66.51% on ICDAR 2011, ICDAR 2013, ICDAR 2015, MSRA-TD500, ReCTS 2019 and Multi-lingual 2019 datasets, respectively. |
topic |
Text detection mobile-network mobile devices multi-oriented text multi-lingual convolutional neural network |
url |
https://ieeexplore.ieee.org/document/9289838/ |
work_keys_str_mv |
AT manuelcordova peleetextx002bx002batinyneuralnetworkforscenetextdetection AT allanpinto peleetextx002bx002batinyneuralnetworkforscenetextdetection AT heliopedrini peleetextx002bx002batinyneuralnetworkforscenetextdetection AT ricardodasilvatorres peleetextx002bx002batinyneuralnetworkforscenetextdetection |
_version_ |
1724182929556373504 |