OCR Data Capture With Background Security Features.

Audience

This article is for support, QA, implementation, and field engineering teams configuring OCR data capture for documents that also use background security features.

Applies To

• Builds that include the split OCR capture stream change.
• Validated with the IPP Class Driver.
• Validated with pdfocr8 OCR capture and no white background cleanup options enabled.

Problem This Solves

When OCR was enabled with pdfocr8, data capture could work, but the OCR output could include a full-page white layer. If a template placed security features behind the document content, that white layer could hide the background features.

The split-stream change fixes this by using:

• pdfwrite for the final printable document stream.
• pdfocr8 for a separate OCR capture stream.

The user-facing configuration can still specify pdfocr8. In the split-stream build, the application internally keeps the printable stream on pdfwrite when a raster OCR engine is used, then sends the OCR stream to data capture.

Recommended OCR Settings

Use the following baseline settings for the validated IPP Class Driver workflow:

{
  "Engine": "pdfocr8",
  "TryOCR": true,
  "OCRMode": "Always",
  "OCRLanguage": "eng",
  "Segment1": " -sColorConversionStrategy=LeaveColorUnchanged -dALLOWPSTRANSPARENCY ",
  "Segment2": "",
  "Segment3": "",
  "CaptureEngine": "",
  "CaptureSegment1": "",
  "CaptureSegment2": "",
  "CaptureSegment3": ""
}

If the OCR language data is not found automatically, also configure the OCR data path:

{
  "OCRDataPath": "C:\\Program Files\\Troy Group\\Troy Sentry Print Client\\Print Processor Service\\Resources\\tessdata"
}

Template Settings

For this workflow, leave white background cleanup options off.

Do not enable:

• Remove white vector fill backgrounds, method 1.
• Remove white vector fill backgrounds, method 2.
• Remove OCR white page background.

Known note: Remove white vector fill backgrounds (method 1) has been observed to throw an error and prevent printing in this scenario. Leave it off unless engineering specifically asks for a controlled test.

Expected Behavior

With the recommended settings:

• Data capture should populate successfully.
• Portrait documents should print in portrait orientation.
• Background security features should remain behind the page content and remain visible.
• Security features should not need to be moved to the foreground just to appear.
• White background cleanup should not be required.

Configuration Notes

• Keep Engine set to pdfocr8 when OCR data capture is required.
• Keep TryOCR set to true.
• Use OCRMode set to Always for documents that require OCR capture.
• Use OCRLanguage set to eng unless another language is required and installed.
• Keep CaptureEngine blank for the validated configuration. Blank allows the application to automatically use the OCR engine for the capture stream while keeping the print stream on pdfwrite.
• Keep the Segment1 value shown above so Ghostscript preserves color behavior and PostScript transparency handling.

Validation Checklist

Before calling the configuration complete, verify:

• The workstation or server is running the split-stream build.
• The printer is using the IPP Class Driver.
• The OCR settings match the recommended baseline.
• White background cleanup options are disabled in the template.
• A test portrait document prints successfully.
• Data capture fields populate correctly.
• Background security features are visible without moving them to the foreground.

Troubleshooting

If data capture does not populate

• Confirm Engine is pdfocr8.
• Confirm TryOCR is true.
• Confirm OCRMode is Always.
• Confirm OCRLanguage is eng.
• Confirm the tessdata files are installed and OCRDataPath is set if needed.
• Confirm the environment is using the split-stream build, not an older build.

If background features are hidden

• Confirm white background cleanup options are off.
• Confirm the template feature is actually configured as a background feature.
• Confirm the installed build contains the split-stream change.
• Confirm the job is being sent through the IPP Class Driver path used during validation.

If printing stops after enabling cleanup

• Turn cleanup back off.
• Retest with the recommended OCR settings.
• Capture the print processor logs and the exact template setting that caused the failure.

If OCR output looks like junk data

• Confirm the source job is being sent through the IPP Class Driver.
• Confirm OCRMode is Always.
• Confirm pdfocr8 is available in the installed Ghostscript/runtime path.
• Confirm the document is not being captured from the printable stream only.