Read Zone

Can be one of the following types:

Specifies the area of the page to extract text from.
The region can be defined as a fixed rectangle, or as an anchored region that moves dynamically based on a detected label or pattern.

• For fixed zones, configure the coordinates to match the location of the desired value on the document.
• For anchored zones, use an anchor pattern to locate a label or keyword, and define the region relative to that anchor.
• Anchored regions are useful for extracting values that appear next to labels, such as "Account Number:" or "Date:".

Example

To extract a value next to the label Invoice Number:, configure the anchor pattern to match the label, and set the region to begin just after the anchor.

> Accurate region definition is critical for reliable extraction. Use the zone editor to visually configure and test the region.

Selects the OCR Profile for extracting text from the region.
If set, the region is processed using OCR; otherwise, text is extracted directly from the document's text layer.

• Use OCR for image-based or scanned documents where text is not natively available.
• The OCR Profile controls recognition settings, preprocessing, and filtering.
• If left blank, extraction will use existing text data (if available).

> Choose an OCR Profile that matches the document type and quality for best results.

Specifies an IP Profile (image processing profile) to apply to the region before OCR is performed.
Use this to improve OCR accuracy by cleaning up noise, adjusting contrast, or correcting skew.

• Only applies when an OCR Profile is set.
• Common IP steps include deskew, binarization, and noise removal.
• Avoid IP commands that change image size or resolution, as this may affect region alignment.

> Use the diagnostic mode to review before-and-after images and tune your IP Profile for optimal OCR results.

Can be one of the following types:

The 'Orientation' property sets the expected rotation (in degrees) of the text within the extraction region.
Use this property to ensure accurate reading of text that is rotated or printed sideways in the document zone.

Purpose

Many documents contain fields or values printed at an angle, such as vertical labels, rotated table columns, or sideways text. The 'Orientation' property allows you to specify the rotation so that extraction (especially OCR) is performed correctly.

Configuration Guidance

• Set to 0 for normal horizontal text.
• Set to 90, 180, or 270 for rotated text, depending on the direction.
• Only applies when an OCR Profile is set; direct text extraction ignores this property.
• If the region contains rotated text, match the orientation to the actual rotation for best results.
• If unsure, use the diagnostic mode to view the region and verify the rotation visually.

Impact

• Correct orientation improves OCR accuracy and ensures the extracted value matches the intended field.
• Incorrect orientation may result in poor recognition, missing values, or garbled output.
• Changing this value does not affect the region's position, only how the text is interpreted during extraction.

Example

If a field is printed vertically on the page, set 'Orientation' to 90 to rotate the region for OCR. If a table column is upside down, set 'Orientation' to 180.

> Use diagnostics to confirm the orientation and troubleshoot extraction issues.

Can be one of the following types:

The 'Value Extractor' property allows you to run a secondary extraction process on the text found in the region.
This enables targeted extraction of specific values, patterns, or entities from the zone's content.

Purpose

Use a Value Extractor to refine the results of zonal extraction, such as isolating a date, number, or keyword from a block of text. This is especially useful when the region contains extra or variable text, and you need to extract only a specific value.

Configuration Guidance

• Assign a Value Extractor such as Pattern Match, List Match, or a custom extractor to this property.
• Configure the nested extractor to match the desired value (e.g., a regular expression for a date, or a list of valid codes).
• If not set, the full text content of the region is returned.
• The nested extractor operates only on the text found in the region, not the entire document.
• Use the 'Discard Misses' property to control what happens if the extractor finds no matches.

Impact

• Enables precise extraction of values from complex or noisy regions.
• Supports chaining extractors for advanced scenarios, such as extracting a value next to a label, then matching a pattern within that value.
• Improves data quality by filtering out unwanted text and focusing on the target value.

Example

To extract a date from a region, assign a Pattern Match extractor with a date regular expression:
\d{2}/\d{2}/\d{4}

To extract a code from a list, assign a List Match extractor with the valid codes.

> Use diagnostics to review the output of the nested extractor and tune its configuration for best results.

Controls the output when the Value Extractor finds no matches in the region.

• If true, no value will be output when the extractor finds no matches.
• If false, the full text of the zone will be output as a fallback.

> This setting is useful for controlling how missing or unexpected values are handled in downstream processes.

When multiple values are found by the Value Extractor, they are combined using this separator.

• Supports special escape sequences:
  • \r = Carriage return
  • \n = Line feed
  • \t = Tab
  • \f = Form feed
  • \s = Space
• Use this property to control how multiple extracted values are joined in the output.
• If left blank, values are output as separate instances.

Example

To join multiple values with a comma, set the separator to ,.

> Choose a separator that matches your downstream data requirements.

Controls how line breaks in the extracted text are represented.

• Supports special escape sequences:
  • \r = Carriage return
  • \n = Line feed
  • \t = Tab
  • \f = Form feed
  • \s = Space
• Use this property to format multi-line output for compatibility with downstream systems.

Example

To replace line breaks with a space, set the separator to \s.

> Adjust this setting to match the expected format for your data consumers.

The 'Output Full Region' property controls how the region (zone) is assigned to each extracted value instance.
This affects how results are visualized, reviewed, and used downstream.

Purpose

By default, the region assigned to each output instance matches the bounds of the actual text found.
When enabled, this property causes the region to reflect the full extraction area (the entire zone), regardless of where text was found within it.

• Set to true to assign the full region (zone) to each output instance.
• Set to false to assign only the bounding box of the extracted text.

Configuration Guidance

• Use true when you want to visualize or debug the entire OCR area, or when downstream processes require the full region.
• Use false for more precise highlighting and to focus on the actual value location.

Impact

• Affects how results are displayed in the UI and in diagnostic images.
• Can help troubleshoot region configuration by showing the full area being processed.
• Does not affect the extracted value itself, only the region metadata.

Example

If extracting a value from a large zone, enabling 'Output Full Region' will highlight the entire zone, even if the value is a small portion of it.

> Use diagnostics to compare the full region and text bounds for validation and tuning.

The 'Exclude Anchor' property enables automatic removal of anchor text from the extracted result when using anchored regions.
This is useful for extracting values that appear next to labels or keywords, without including the label itself.

Purpose

Anchored regions are defined relative to a detected label or pattern (the anchor).
When enabled, this property removes the anchor text from the output, so only the value of interest is returned.

Configuration Guidance

• Only applies to anchored regions (where the 'Location' property is anchored).
• Enable to exclude label text such as Account Number: or Date: from the extracted value.
• If the anchor is matched fuzzily (e.g., due to OCR errors), some misrecognized characters may remain in the output.
• Use diagnostics to review anchor exclusion and ensure the output matches expectations.

Impact

• Improves data quality by removing unwanted label or context text.
• Essential for scenarios where the value is adjacent to a label and only the value is needed.
• May require tuning of the anchor pattern for best results.

Example

To extract the value next to Invoice Number: and exclude the label, enable this property.

> Use diagnostics to validate anchor exclusion and troubleshoot any remaining label text.