Skip to main content


Deep Learning and Neural Networks have been a driving force behind Optical Character Recognition. The idea behind Neural Networks is to make a program recognize characters as a human brain would with a high accuracy rate, however for the most part this is dealt with the English alphabet and numbers. Currently, Mandarin Chinese is the most spoken language in the world and the characters are the most difficult to recognize. This is because every word has a different character which is almost represented as a picture, so a program that could convert these strokes and Chinese characters into binary code would be extremely beneficial to the general population. This research paper focuses on how Neural Networks can be used to help better understand current technology such as Google Translate and Pleco by taking handwritten characters and allowing the computer to recognize them. For the implementation part of this paper, the focus is on Optical Character Recognition for Chinese handwritten characters from the CASIA Handwritten Database and using the open source software library Tensorflow.


This is a metadata-only record.



  • Subject
    • Computer Science & Information Systems

  • Institution
    • Dahlonega

  • Event date
    • 25 March 2016

  • Date submitted

    18 July 2022

  • Additional information
    • Acknowledgements:

      Dr. Bryson Payne