Home > About Us > Technology > Technical Report > Technical Report No.25 2016

System to Digitize Information Written on Forms

By using image recognition technology and character recognition technology accumulated over many years, Fuji Xerox has developed a form data entry system called FormDataEntry that can recognize and digitize characters and marks written on paper forms with high precision. By using this system, it is possible to quickly sort and digitize information written on forms created in various types of businesses and work. The system can be used for forms directly exchanged with customers or users over the counter in financial institutions and public services, forms mailed for filing applications and conducting surveys, forms faxed for sending and receiving orders, etc. With character recognition technology that uses semi-Marcov conditional neural fields and single character recognition technology that uses convolutional neural networks, the system is capable of recognizing handwritten characters with high precision. This paper describes the system configuration and the basic technologies used for the system.

Authors

  • Toshiya Koyama
    Incubation Center, Research & Technology Group
  • Shinichi Miyamoto
    Incubation Center, Research & Technology Group
  • Shinichi Yada
    Incubation Center, Research & Technology Group
  • Shunichi Kimura
    Incubation Center, Research & Technology Group
  • Takaaki Sakakibara
    Incubation Center, Research & Technology Group