Welcome to Kofax RPA > Robot Building > Robot Steps > Extract Text From Image > Extended OCR Settings

Extended OCR Settings

Kofax RPA provides optical character recognition (OCR) functionality to extract text from images and to automate applications with limited or no automation API.

OCR is a complicated process and recognition results depend on many factors, such as screen fonts, background and foreground color, text size, and so on. Kofax RPA installs the ocr.cfg file that contains some configuration settings you can use to alter recognition results. The file includes detailed description of configuration settings. The ocr.cfg file is located in the Kofax RPA installation directory as follows.

  • On the Windows-based automated computer with installed Desktop Automation Service:

    DesktopAutomationService\lib in the Desktop Automation service installation directory. Example:

    C:\Program Files\Kofax RPA DesktopAutomation 11.2.0.0 \DesktopAutomationService\lib

  • On the local Windows-based computer to use with the built-in browser:

    nativelib\hub\windows-x64\<build number>\lib* in the Kofax RPA installation directory. Example:

    C:\Program Files\Kofax RPA 11.2.0\nativelib\hub\windows-x64\166\lib

  • On the local Linux-based computer to use with the built-in browser:

    nativelib/hub/linux-x64/<build number>/lib in the Kofax RPA installation directory. Example:

    Kofax RPA_11.2.0.0/nativelib/hub/linux-x64/166/lib

* The build number is different in different versions of the program.

Change OCR engine and language

Change OCR engine

Kofax RPA uses either Tesseract (default) or OmniPage engine to capture text from images. To change the OCR engine from default, perform the following steps.

  1. Locate the ocr.cfg file on your computer.

  2. Open ocr.cfg in a text editor and locate the engine_type option.

  3. Specify an OCR engine, such as omnipage as a value as follows engine_type = omnipage.

    If you want to use the default OCR engine (Tesseract), either specify tesseract as a value in the engine_type option or just delete any value from this option.

Change OCR language

  1. Locate the ocr.cfg file on your computer.

  2. In the text editor, open ocr.cfg and locate the default_language option.
  3. Either replace eng with another language code, for example jpn or, if you want to use more than one language, add jpn using the plus sign, such as default_language=eng+jpn. The language code must be in ISO 639-3 or ISO 639-1 format. Save and close the file.

OmniPage includes all supported languages in the installation. For Tesseract, only English language is included in the installation. To add more languages for UI recognition by Tesseract, see step 1 and 2 in the "Change or add UI recognition language for Tesseract" section in Tree Modes.

Important OCR engine and language settings for the Desktop Automation Service are specified in the Desktop Automation Service configuration window separately for each computer running the service. See Configure Desktop Automation Service for details.

Image preprocessing

The following information is applicable to Tesseract engine only.

Before the actual OCR process is initiated for an image, the image is preprocessed using a particular algorithm. In the ocr.cfg file, the preparation setting defines the algorithm to use. By default, it is set to normal.

Note In case the default preprocessing algorithm gives a result that you find unsatisfactory, you may try switching to a different algorithm. To do so, change the value of preparation to 10.2 and save the changes.