8.0

4.3. Creating a document template

The most important step in setting up a project is the creation of a template. The quality of data received

after forms have been processed depends on the correctness of the template. To create a template, you

must specify the:

• Static elements on the image: anchors, separators, static text, and barcodes. Select which of these

elements are to be used for template matching and document identification. Anchors are detected

and marked automatically.

• Location of all fields. Fields must correspond to the areas of the image from which data is to be

extracted.

• Properties of each field: select the data types to be searched for in every field (this significantly

improves the recognition quality) specify which fields must be sent to the operator for

verification, etc.

• Rules by which field values are to be checked. Such rules help the program detect documents

whose values do not conform to certain conditions, for example where a field value does not

correspond to the values of the necessary database.

• Method of data export. Data can be exported to a file or database, or in accordance with a script

procedure.

Once the template is created, it must be published in order to use it for subsequent document

recognition.

To create a new template, from the menu select Project > Document Templates… Click New… in the

dialog box that opens. This will start the Document Template Creation Wizard. In the Create New

Document Template dialog box you can specify the template’s main properties: its name, description,

locale, and writing style. Select the text type: ICR (hand-printed) if most document fields are filled out

by hand, or OCR (printed) if the values in most document fields are printed. In the latter case, select the

print type from the drop-down list. The text type specified at this stage will be the default text type, but

you will be able to change the text type for individual fields.

Next, load or scan the image on the basis of which the template is created. If your document consists of

several pages, load the first page. When adding the rest of the pages, please refer to the

recommendations of the section Creating a template for a multi-page document

. You can scan the image

of a blank page or load it from a file. If you are going to process semi-structured documents, you must

use a flexible description when creating a template. If this is the case, select the Load FlexiLayout

option and specify the path to the AFL file containing the flexible description created in ABBYY

FlexiLayout Studio.

You can now select the field types that are to be automatically detected. You can specify checkmarks

and text entry fields. Searching automatically for text fields with marking and rectangular checkmarks is

highly efficient. However, if text fields on your form contain no marking and checkmarks need to be

made against a plain white background, we recommend that you mark such fields manually.

If there are anchors on the image, these will be detected and marked automatically.