Result Filter

Use the 'Allowed Pages' property to restrict results to only those found on specific pages of the document.

• Specify a comma-separated list or range (e.g., 1,3,5-7) to include only those pages.
• If left blank, results from all pages are included.

This is useful for scenarios where only certain pages are relevant for extraction, such as cover pages, headers, or footers.

Set the 'Minimum Confidence' property to exclude results with a confidence score below the specified threshold.

• Confidence is a value between 0 and 1, where higher values indicate greater certainty in the extraction.
• Results with a lower confidence than this value are discarded.

Use this to improve data quality by filtering out uncertain or low-confidence matches.

The 'Length Range' property allows you to specify minimum and/or maximum character counts for results.

• Only results whose value length falls within this range are included.
• Leave blank to allow any length.

This is useful for excluding results that are too short (e.g., noise) or too long (e.g., paragraphs when only short codes are desired).

Use 'Character Classes' to require that results match specific character composition rules.

• Each Character Class Filter can enforce requirements such as "at least 80% numeric" or "no more than 2 punctuation characters".
• Multiple filters can be combined; all must be satisfied for a result to be included.

This is useful for targeting results like numeric codes, uppercase words, or values with specific formatting.

Can be one of the following types:

The 'Exclusions' property allows you to specify a Value Extractor whose results are used to exclude overlapping matches.

• Any result that overlaps with an exclusion match is discarded.
• This is useful for removing known false positives, headers, footers, or other unwanted content.

Use this to refine the result set by removing results that conflict with or duplicate other extracted values.

The 'Region' property restricts results to those whose geometric center (or other criteria, see 'ROIMode') falls within a specified rectangle.

• Specify a rectangle to define the area of interest.
• Leave blank (0,0):(0,0) to disable region filtering.

This is useful for extracting values from specific zones, such as headers, footers, or form fields.

Can be one of the following values:

The 'ROIMode' property determines the inclusion logic for region-based filtering:

• Center: Only results whose center point falls inside the region are included.
• Intersect: Results that intersect the region in any way are included.
• Contains: Only results fully contained within the region are included.

Adjust this setting to control how strictly results must match the specified region.

The 'Width Range' property allows you to specify minimum and/or maximum geometric widths for results.

• Only results whose width falls within this range are included.
• Leave blank to allow any width.

Use this to target results of a specific size, such as wide headings or narrow codes.

The 'Height Range' property allows you to specify minimum and/or maximum geometric heights for results.

• Only results whose height falls within this range are included.
• Leave blank to allow any height.

Use this to target results of a specific vertical size, such as tall titles or short codes.

The 'Font Size Range' property allows you to specify minimum and/or maximum font sizes for results.

• Only results whose font size (in points) falls within this range are included.
• Leave blank to allow any font size.

Use this to target results with specific typographic emphasis, such as large headings or small footnotes.

Can be one of the following types:

The 'Horizontal Alignment' property allows you to require that results have a specific alignment (left, center, or right) relative to the page.

• If not set, results of any alignment are included.

Use this to isolate results that are consistently aligned, such as columnar data or centered titles.

A combination of the following flags:

The 'Character Casing' property allows you to require that results match a specific casing style (e.g., uppercase, lowercase, title case).

• Select one or more casing styles to allow.
• Use in combination with 'Case Similarity' to control how closely results must match the desired casing.

This is useful for targeting headings, acronyms, or other elements with distinctive casing.

The 'Case Similarity' property sets the minimum percentage of characters that must match the selected casing style.

• Value is between 0 and 1 (e.g., 0.8 = 80%).
• Higher values require stricter adherence to the selected casing.

Use this to fine-tune casing requirements, allowing for minor deviations or enforcing strict casing as needed.

The 'Minimum Font Scale' property allows you to require that results have a font size at least a certain percentage larger than the document's median font size.

• Value is a percentage (e.g., 0.2 = 20% larger).
• Use this to target large print elements such as document titles or section headings.

This is useful for isolating visually prominent elements that stand out from the main body text.

The 'Require Underline' property allows you to include only results that are visually underlined in the document.

• When enabled, only underlined results are included.
• Use this to target headings, labels, or other elements that are consistently underlined for emphasis.

This is useful for extracting section headers, form fields, or other visually distinct elements.