How does Scanbot Text Recognition (OCR) work?

Bennet Conrads -

Scanbot Pro OCR: How it works and what you need to know in order to use Scanbot's Text Recognition Feature

 

First off, here are some basic information about our OCR technology:

OCR stands for optical character recognition. Our cutting-edge OCR technology recognizes and extracts the text from your scanned documents.

Here are some basic info about our OCR engine:

  • OCR is only performed locally on your device, i.e. we don't and won't send anything outside to an unknown server
  • Currently Scanbot supports OCR for over 60 different languages.
  • We've combined several technologies for OCR engine.

 

So how does it work?

A pre-condition to use OCR is to be a Scanbot Pro user. After purchasing Scanbot Pro, your settings will offer the additional Pro features. Hit Text Recognition (OCR) to enter the configurations. The top row defines whether Text Recognition is enabled or not. Below, you can find the list of supported OCR languages. Make sure that the language(s) you need are enabled, but please note that the OCR result is better when when you select less languages.

 

 Activate all languages you want to detect in the settings

Now that OCR has been activated, each scan you will perform will have OCR applied to it. The first time you scan a document you will see a short message saying 'Waiting for OCR files...'. In this step Scanbot is initializing OCR. The very first time, this may take a little longer as Scanbot needs to download essential support files for your language selection.

 

Next, you will see the message 'Processing Pages...'. In this step, OCR is actually applied to your scan. Depending on the scan quality and the content of the scan, this may take (in rare cases) 2-3 minutes, but usually only 30-40 seconds. The progress is indicated in percent.

 

Now OCR is done. To access the contents just hit the scan in the list view, which brings you to the document view. Here tap Text at the top (just below the file name)

 

Open the detail view to see your document  Switch to the text tab to see the recognized text

 

The extracted text will now be shown separately and can be shared and searched.
If the document contains text such as phone numbers, emails or addresses you can use them with one tap in the Actions Menu.

 

Moreover, as soon as OCR has been applied to a PDF scan, all contents are searchable by using Spotlight Search and the Scanbot search.

 

A few more tips: To ensure the best OCR results, please consider the following:

  • The better the document quality the better the scan and therefore the OCR result.
  • Make sure that your scan gets enough light.
  • Ensure that the languages are correctly configured in the OCR configurator.
  • Please also make sure that the image orientation is in portrait mode and not in landscape mode. Currently, our OCR engine can only work with portrait scans.

 

Manual OCR

 

Still have some old documents lying around that have not been OCR’d yet?


Here’s how you can apply the text recognition manually:

  1. Open the document
  2. Tap Text at the top
  3. Tap Re-Run OCR below the message that no text was found
Was this article helpful?
1 out of 1 found this helpful
Have more questions? Submit a request