Set up an OCR skill - ABBYY Documentation

Create an OCR skill in the Skill Catalog, and then configure it in the Create OCR Skill dialog box.

Create and configure an OCR skill

Create a new OCR skill

In the Skill Catalog, click Create in the toolbar.

Select the OCR skill type

Select the OCR Skill skill type. The Create OCR Skill dialog box opens.

Configure General settings

On the General tab:

Enter a name and description for the new skill.
Select the Technology Core version, preferably the one marked (recommended).

Under Image Processing, select a PDF Processing Mode for PDF documents:

Mode	What it does
Default (Recommended)	Default setting. Uses the embedded PDF text layer when available and supplements it with OCR as needed.
Use OCR Only	Ignores any embedded PDF text layer and performs full OCR on the document.
Use Text Layer Only	Extracts text from the embedded PDF text layer. If no text layer exists, Vantage falls back to OCR automatically.

For more information, see PDF Processing Mode.

Select recognition languages

On the Languages tab:

In the Allowed Languages section, select one or more document recognition languages. During processing, Vantage automatically chooses the document language from the selected languages. The number of selected languages can affect recognition speed and quality.
If the document contains handwritten text, enable the Handwritten option in the Text Appearance section.

Configure image enhancements

On the Image Enhancements tab, Crop Image and Correct page orientation are enabled by default. Turn off either option if your documents do not require it.

Option	What it does
Crop Image	Crops the image to the edges of the original document.
Correct page orientation	Rotates the image automatically to restore its original orientation.

Enable barcode recognition

On the Barcodes tab, enable the barcode types that may appear on your documents. The number of selected barcode types can affect recognition speed.If you don’t need to recognize any barcodes, click the button in the Barcode Types table header that shows the number of selected options. Then click Deselect all.

Choose export formats

On the Export tab, select one or more document export formats. For the options available for each format, see Available export formats.

If a transaction contains several files, the OCR skill merges them into one document. The transaction produces one output file per selected export format.

Create and publish the skill

Click Create and Publish.

Available export formats

JSON is the default export format. The Export tab offers the following formats and options.

Format	Option	Description
JSON	Text only	Default. A JSON file that contains only recognized text, without the document layout.
JSON	Preserve document structure	A JSON file that contains recognized text and preserves the document layout.
XML	Text only	An XML file that contains only recognized text, without the document layout.
XML	Preserve document structure	An XML file that contains recognized text and preserves the document layout.
ALTOXML	Text only	An ALTO XML file that contains only recognized text, without the document layout.
ALTOXML	Preserve document structure	An ALTO XML file that contains recognized text and preserves the document layout.
TXT		A plain text file.
HTML		An HTML document.
DOCX	Editable	Default. An editable Microsoft Word document that preserves the original text flow. The output may not look exactly like the original.
DOCX	Exact	A non-editable Microsoft Word document that looks exactly like the original.
XLSX		A Microsoft Excel workbook.
PPTX		A Microsoft PowerPoint presentation.
PDF	PDF/A-3a (smaller file size)	Default. An editable text-under-image PDF in the PDF/A-3a standard, with a smaller file size.
PDF	PDF/A-3a (maximum quality)	An editable text-under-image PDF in the PDF/A-3a standard, with maximum image quality.
PDF	PDF/A-3b (smaller file size)	An editable text-under-image PDF in the PDF/A-3b standard, with a smaller file size.
PDF	PDF/A-3b (maximum quality)	An editable text-under-image PDF in the PDF/A-3b standard, with maximum image quality.
PDF	Image-only	A non-editable image-only PDF in the PDF/A-3b standard.
TIFF		A TIFF image.
JPEG	Maximum quality	Default. An enhanced JPEG image at 95% quality.
JPEG	Smaller file size	An enhanced JPEG image at 75% quality, with a smaller file size.

Selecting Text only for JSON, XML, or ALTOXML makes export to DOCX, XLSX, and PPTX unavailable. JSON, XML, and ALTOXML always use the same option. If you change the option for one of these formats, Vantage applies it to the other two. The smaller file size PDF options use Mixed Raster Content (MRC) compression, which sets optimal compression rates separately for the text, the pictures, and the background.

​Create and configure an OCR skill

​Available export formats

​Related topics

Create and configure an OCR skill

Available export formats

Related topics