Perform OCR on multilingual documents

Optical character recognition (OCR) enables you to locate and extract text from images or the screen.

Although most scenarios require you to handle text in a specific language, there are cases where the sources are multilingual.

To perform OCR on these sources, use a Tesseract engine in the respective OCR action and enable the Use other languages option in the engine settings.

The Use other languages option in the Exctract text witg OCR action.

When the Use other languages option is enabled, the action displays two additional settings: the Language abbreviation and Language data path fields.

The Language abbreviation field indicates to the engine which language to look for during OCR. The Language data path field contains the language data files (.traineddata) used to train the OCR engine.

The Language abbreviation and Language data path fields in the Exctract text witg OCR action.

After downloading the data files for the needed languages, move them to a common folder to make them available under the same path.

Next, select the created folder in the Language data path field, and populate the corresponding language codes in the Language abbreviation field. To separate the language codes, use the plus character (+).

Note

You can find all the available language codes in the source of the language data files. In the following example, the used codes represent Telugu, Hindi, and English.

The populated Language abbreviation and Language data path fields in the Exctract text witg OCR action.