View in EDS

Fast multi-language LSTM-based online handwriting recognition.

Saved in:
Bibliographic Details
Title: Fast multi-language LSTM-based online handwriting recognition.
Authors: Carbune, Victor1 vcarbune@google.com, Gonnet, Pedro1, Deselaers, Thomas1, Rowley, Henry A.2, Daryin, Alexander1, Calvo, Marcos1, Wang, Li-Lun2, Keysers, Daniel1, Feuz, Sandro1, Gervais, Philippe1
Source: International Journal on Document Analysis & Recognition. Jun2020, Vol. 23 Issue 2, p89-102. 14p.
Subjects: Graphology, Deep learning, Language & languages, Artificial neural networks, Error rates
Abstract: We describe an online handwriting system that is able to support 102 languages using a deep neural network architecture. This new system has completely replaced our previous segment-and-decode-based system and reduced the error rate by 20–40% relative for most languages. Further, we report new state-of-the-art results on IAM-OnDB for both the open and closed dataset setting. The system combines methods from sequence recognition with a new input encoding using Bézier curves. This leads to up to 10 × faster recognition times compared to our previous system. Through a series of experiments, we determine the optimal configuration of our models and report the results of our setup on a number of additional public datasets. [ABSTRACT FROM AUTHOR]
Copyright of International Journal on Document Analysis & Recognition is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
FullText Text:
  Availability: 0
Header DbId: egs
DbLabel: Engineering Source
An: 143359731
AccessLevel: 6
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Fast multi-language LSTM-based online handwriting recognition.
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Carbune%2C+Victor%22">Carbune, Victor</searchLink><relatesTo>1</relatesTo><i> vcarbune@google.com</i><br /><searchLink fieldCode="AR" term="%22Gonnet%2C+Pedro%22">Gonnet, Pedro</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Deselaers%2C+Thomas%22">Deselaers, Thomas</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Rowley%2C+Henry+A%2E%22">Rowley, Henry A.</searchLink><relatesTo>2</relatesTo><br /><searchLink fieldCode="AR" term="%22Daryin%2C+Alexander%22">Daryin, Alexander</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Calvo%2C+Marcos%22">Calvo, Marcos</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Wang%2C+Li-Lun%22">Wang, Li-Lun</searchLink><relatesTo>2</relatesTo><br /><searchLink fieldCode="AR" term="%22Keysers%2C+Daniel%22">Keysers, Daniel</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Feuz%2C+Sandro%22">Feuz, Sandro</searchLink><relatesTo>1</relatesTo><br /><searchLink fieldCode="AR" term="%22Gervais%2C+Philippe%22">Gervais, Philippe</searchLink><relatesTo>1</relatesTo>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="JN" term="%22International+Journal+on+Document+Analysis+%26+Recognition%22">International Journal on Document Analysis & Recognition</searchLink>. Jun2020, Vol. 23 Issue 2, p89-102. 14p.
– Name: Subject
  Label: Subjects
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Graphology%22">Graphology</searchLink><br /><searchLink fieldCode="DE" term="%22Deep+learning%22">Deep learning</searchLink><br /><searchLink fieldCode="DE" term="%22Language+%26+languages%22">Language & languages</searchLink><br /><searchLink fieldCode="DE" term="%22Artificial+neural+networks%22">Artificial neural networks</searchLink><br /><searchLink fieldCode="DE" term="%22Error+rates%22">Error rates</searchLink>
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: We describe an online handwriting system that is able to support 102 languages using a deep neural network architecture. This new system has completely replaced our previous segment-and-decode-based system and reduced the error rate by 20–40% relative for most languages. Further, we report new state-of-the-art results on IAM-OnDB for both the open and closed dataset setting. The system combines methods from sequence recognition with a new input encoding using Bézier curves. This leads to up to 10 × faster recognition times compared to our previous system. Through a series of experiments, we determine the optimal configuration of our models and report the results of our setup on a number of additional public datasets. [ABSTRACT FROM AUTHOR]
– Name: AbstractSuppliedCopyright
  Label:
  Group: Ab
  Data: <i>Copyright of International Journal on Document Analysis & Recognition is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract.</i> (Copyright applies to all Abstracts.)
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=egs&AN=143359731
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1007/s10032-020-00350-4
    Languages:
      – Code: eng
        Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 14
        StartPage: 89
    Subjects:
      – SubjectFull: Graphology
        Type: general
      – SubjectFull: Deep learning
        Type: general
      – SubjectFull: Language & languages
        Type: general
      – SubjectFull: Artificial neural networks
        Type: general
      – SubjectFull: Error rates
        Type: general
    Titles:
      – TitleFull: Fast multi-language LSTM-based online handwriting recognition.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Carbune, Victor
      – PersonEntity:
          Name:
            NameFull: Gonnet, Pedro
      – PersonEntity:
          Name:
            NameFull: Deselaers, Thomas
      – PersonEntity:
          Name:
            NameFull: Rowley, Henry A.
      – PersonEntity:
          Name:
            NameFull: Daryin, Alexander
      – PersonEntity:
          Name:
            NameFull: Calvo, Marcos
      – PersonEntity:
          Name:
            NameFull: Wang, Li-Lun
      – PersonEntity:
          Name:
            NameFull: Keysers, Daniel
      – PersonEntity:
          Name:
            NameFull: Feuz, Sandro
      – PersonEntity:
          Name:
            NameFull: Gervais, Philippe
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 06
              Text: Jun2020
              Type: published
              Y: 2020
          Identifiers:
            – Type: issn-print
              Value: 14332833
          Numbering:
            – Type: volume
              Value: 23
            – Type: issue
              Value: 2
          Titles:
            – TitleFull: International Journal on Document Analysis & Recognition
              Type: main
ResultId 1