Skip to main content
Version: V2.9

Text Recognition

Business scenario description

General text recognition can recognize text content on common documents. such as:

  • It can recognize unstructured long text documents such as contracts, legal documents, and red-head documents, and directly archive the recognized text. Or after recognition, use customized AI capability information extraction to structure the recognized long text.
  • It can recognize semi-structured documents such as delivery note, overseas invoice, and insurance policy, and then use customized AI capability, which is custom template, to structure the documents based on position and text information.

Converting offline data into online data with general text recognition can free the staff from repetitive typing tasks and devote themselves to more high-leverage work.

Features

General text recognition has the following features:

  • Multi-language:Support more than 20 common languages such as: Chinese (Traditional and Simplified), English, Japanese, Korean, French, German, Portuguese, Italian, Spanish, Dutch, etc.
  • Complex scenarios:It can deal with complex scenarios such as rotation, occlusion, tilt, dense text, complex background, uneven lighting, blurring, and handwriting.
  • Multiple formats:It supports jpeg, jpg, png, pdf, bmp and tiff format.
  • High accuracy:The accuracy of printed body recognition can reach 99%.

Instructions

Text version

1)Log in to the Mage platform and click the menu of Text Recognition. 通用文字1

2)Create a new general text recognition model. 通用文字2

3)Click the tester in the upper right corner to enter the test page. 通用文字3

4)Upload a document, click the button “start the test”, you will get the recognition result. 通用文字4