Capture Server Pro - Tips for Adding Identification Zones for Classification in Assist

Identification zones are used for classifying documents and serving as a reference point for OCR zones. You can also use identification zones to set static metadata in the document, such as a property that will always have a particular value when certain features, like VOID stamps or logos, are present on the page.

When you place an identification zone, you can apply the Classify interaction, which marks the zone for use in determining which documents match the template. If you select multiple identification zones, you can link them together, allowing you to create more complex classification criteria.

Tips for Identification Zones

When a document is processed by the Classify Your Documents activity, it is compared against the configured Assist templates. Each document is scanned to locate areas that match the images in the template’s identification zones that are configured to use the classify interaction.

The end goal of using identification zones to classify documents is to highlight areas on a form that make it different from other forms. To improve template matching, observe the following tips when adding identification zones.

Zones Should Be Visually Distinct

Identification zones do not use optical character recognition (OCR). Instead of reading the text in the zone, Assist instead compares the visual features of the zone to find it within a document. For that reason, you should use areas with distinct visual characteristics, such as company logos, rather than text.

In the example above, the company logo is a better zone to use for classification than the mailing information. Even though the text is specific, the general shape of the text may falsely match other similar text blocks.

Use Multiple Zones

If you receive multiple documents from the same company, some elements may be common across multiple forms. In those cases, you can increase the specificity of your classification by including multiple zones that differentiate similar forms.

In the example above, each zone individually might not be enough to make the form unique. for example, the company logo may appear on multiple different forms from the same company, and the “Invoice” heading may be used on different invoices from different companies. The combination of all three zones, however, is a unique configuration that sets this form apart.