OCR/BCR Extraction Profiles

Owned by Brian Banet
Last updated: May 22, 2023
3 min read

Customers can now modify OCR and BCR extraction profiles directly in the browser.

Extraction is now presented as an option in the GlobalCapture admin page. By default, two extraction profiles will exist (one for Zonal and one for conversion) which is consistent with how the product has always shipped.

For customers without sophisticated capture needs, it may be beneficial to use only a single profile for both Conversion (Full Page OCR) and Zonal OCR. Customers can provision as many extraction profiles as they need.

Creating Profiles

Administrators can create new profiles (1 in the illustration above) or may upload an existing textocr.cfg or fullpageocr.cfg file (2 in the illustration above). If you have never modified settings of the OCR configuration files in your environment, there is no reason to upload them in 3.0. Existing configuration will continue to function when no extraction profiles are set/selected.

Selecting Profiles

Setting a profile to the DEFAULT behavior for OCR or Conversion will automatically use that profile as part of your workflow. The design surface now offers the ability to set the extraction profile on a more granular level. 

For OCR profiles, users will be able to set the profile on the Node (Classify and Convert), Zone, and Workflow level. 

BCR Advanced profiles can be set on the Node (Separation and Delete Pages) as well as the Zone.

Node Extraction

The more granular setting will always over rule the any more general settings.  Order of priority is as follows: Zone → Node → Workflow → Default Profile

Performing OCR on a zone level can be a costly operation from a performance stand point.  When selected, OCR is taking place live as the document is running through the template rule not all at once, preemptively, as normal. It is recommended, to achieve optimal performance, to minimize per zone extraction in favor or system or workflow defaults to those settings.

Technical Notes

  • Extraction profiles are stored in a folder named Extraction in the GlobalCapture base path. For standard deployments, this path is typically “c:\getsmart\captureprocessing”. Profiles are named by their GUID, which can also be found in the GlobalCapture UI, represented by the ID in the profile’s options menu.

  • Below the Extraction folder, a sub-folder named Samples should always be present, and should include a file named “Blank.tif”. This file is used to test the validity of changes made to profiles and MUST always exist in the path.

  • When editing advanced settings, it’s important note that invalid entires will cause processing errors. GlobalCapture makes an effort to ensure user provided settings are valid BEFORE allowing settings to be saved. If OCR settings are manually modified to be invalid, know that runtime processing errors can occur.