Application 1: Automatic Recognition for handwritten characters

◆ Clients

  • Billing Agency

◆ Problems

  • Have to check thousands of billings (a huge number) per day visually
    and want to save time and effort.
  • In addition, there are many items in billings need to check and the mutual relations
    among items are so complicated, which causes human errors.
  • For example, whether the name in the application, the signature by hand,
    and the sealed name are the same.

◆ Solutions

    Dates, handwritten addresses and names on billings can be identified by using AI OCR.

  • For handwritten addresses, AI OCR can check if the addresses are the real place names
    based on the past results.
  • For signatures and seals that are difficult to recognize as characters, AI OCR helps to
    improve the recognition accuracy based on the past results.
  • In addition, AI OCR can judge the validity of application contents from the relevance
    among items on applications.