Text recognition

This feature available in the 4.10.0 beta version only.

The easiest way to enter data for Libraries related to documents is to use Autofill by Image. This feature allows you to scan an image from the file system or made by a camera to find all keywords and save data based on them.

To adjust Autofill by Image, follow the steps below:

  • Go to Autofill tab and click Mobile: Add button icon
  • Choose By Image option there
  • Define whether you want to keep scanned images after text recognition or not and if yes, choose the Field to associate with this image or create a new one
  • Tap Mobile: Add button  icon to Add rule
  • Set a Field that will be filled with recognized data
  • Choose how to set a Field finder
    • If Manually option is chosen, you will need to pick the correct text from the uploaded image. By default, the text will be divided by blocks, but you can split the block or unite a few ones using the correspondent icons under the image.
    • If Anchor Base option is chosen, you will need to set an Anchor Text — a keyword to search for — and an Anchor Position — the position of data that will be used to fill an entry. Also, you can set additional settings to Ignore case (set by default), to Ignore non-word characters (set by default) and to define the number of Permissible errors (one by default). This option is good to fill dates, sums, applicants, etc.
    • If RegEx option is chosen, you will need to define the Regular expression to use as a search algorithm. If the recognized block of text matches the regular expression, it will be used for the field. While setting the regular expression, you can test it using the button below.
    • If Script option is chosen, you can write a function that will be used to search as a search algorithm. The function has two parameters: rect is used as a coordinate where the text starts and text is used for the text itself. If the function returns true, then this block is used as a field. While setting the script, you can text it using the button below.
  • Define the settings of the Field parser:
    • By Default, the app tries to transform the text to the field using the field types. One line is used for one field, but you can check Multi-line value if you need to parse a few lines of text to a single field. Otherwise, you can write Script that will be used to transform a text.
  • Check the entered rule and tap the check icon at the top-left to save it.
  • Add more rules and save all of them by tapping the check icon at the top-left again.

Please note that the text recognition depends on the quality of the image. It is better to scan high-resolution sharp photos.

While filling this Library with images, you will see the scanning process and the highlighted blocks that are used for the data.

Was this article helpful?

Related Articles