Advertisement

Research on Influence of Image Preprocessing on Handwritten Number Recognition Accuracy

  • Tieming ChenEmail author
  • Guangyuan Fu
  • Hongqiao Wang
  • Yuan Li
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 905)

Abstract

In the process of handwritten number recognition, image pretreatment is a key step that has a great influence on the recognition accuracy. By unifying the standard, handwritten digital images are normalized, which can improve the adaptability of handwritten digital recognition algorithms to different writing habits. This article mainly considers the four characteristics of the angle, position, size and strength when writing characters, and how these factors influence four classical handwritten recognition algorithms. According to the four characteristics, tilt correction, offset correction, size normalization and thinning preprocessing were performed one by one to observe the changes of recognition accuracy in four classical algorithms. Through experiments, it is found that the recognition accuracy of the original data set and the scrambling data set are both greatly improved after preprocessing operation. In conclusion, it is resultful to increase the recognition accuracy by image preprocessing in handwritten digital recognition.

Keywords

Image preprocessing Handwritten number recognition CNN Pretreatment 

References

  1. 1.
    Zhang, M., Yu, Z.Q., Yao, S.W.: Image pretreatment research in recognition of handwritten numerals. Microcomput. Inf. 22(16), 256–258 (2006). (in Chinese)Google Scholar
  2. 2.
    Chen, H., Guo, H., Liu, D.Q., Zhang, J.Q.: Handwriting digital recognition system based on tensorflow. Inf. Commun. 3 (2018). (in Chinese)Google Scholar
  3. 3.
    Zhang, Y.S.: License plate recognition key technology related algorithms research and implementation. North Minzu University (2017). (in Chinese)Google Scholar
  4. 4.
    Zhang, C.F., Yang, G.W.: Research on image pretreatment technique in recognition of handwritten number. Comput. Sci. Appl. 6(6), 329–332 (2016). (in Chinese)Google Scholar
  5. 5.
    Huang, Q.Q.: Research on handwritten numeral recognition system based on BP neural network. Central China Normal University (2009). (in Chinese)Google Scholar
  6. 6.
    Zhang, S.H.: Study and realization of algorithms for chinese characters image’s preprocessing. Microcomput. Dev. 13(4), 53–55 (2003). (in Chinese)Google Scholar
  7. 7.
    Zhang, C.F., Yang, G.W., Yue, M.M.: Improving of Zhang parallel thinning algorithm. Inf. Technol. Informatiz. 6, 69–71 (2016). (in Chinese)Google Scholar

Copyright information

© Springer Nature Switzerland AG 2020

Authors and Affiliations

  • Tieming Chen
    • 1
    Email author
  • Guangyuan Fu
    • 1
  • Hongqiao Wang
    • 1
  • Yuan Li
    • 1
  1. 1.Xi’an Research Institute of High-TechnologyXi’anChina

Personalised recommendations