Advertisement

Conclusion and Future Work

  • Saad Bin AhmedEmail author
  • Muhammad Imran Razzak
  • Rubiyah Yusof
Chapter

Abstract

This book presents substantial contribution in the field of document image analysis specifically in cursive scene text recognition. The complexities of Arabic script have been highlighted, and potential of RNN is exploited because of its context learning ability. The RNN-based statistical models were adapted to support the learning of complex Arabic script. In cursive scripts, there are numerous works presented for Chinese and Japanese characters but Arabic script has not exposed to state-of-the-art research methods. Few efforts are reported to date, which obviously are not enough to address the complexities. The scene text recognition can be categorized as a subproblem of Optical Character Recognition (OCR). Although the appearance of samples do not share similarities. Likewise, there is much difference between handwritten and printed Arabic text. The printed text is more clean and clear whereas handwritten text is dominated by writing styles along implicit noise attached with scanned document.

Copyright information

© Springer Nature Singapore Pte Ltd. 2020

Authors and Affiliations

  • Saad Bin Ahmed
    • 1
    • 3
    Email author
  • Muhammad Imran Razzak
    • 2
  • Rubiyah Yusof
    • 3
  1. 1.King Saud bin Abdulaziz University for Health SciencesRiyadhSaudi Arabia
  2. 2.School of Information TechnologyDeakin UniversityGeelongAustralia
  3. 3.Malaysia-Japan International Institute of Technology (M-JIIT)University of Technology MalaysiaKuala LumpurMalaysia

Personalised recommendations