OCR Data Aligner

Coordinates alignment and correction of OCR results between Azure Document Intelligence and Grooper's OCR engines.

Remarks

OCRDataAligner manages the process of aligning, supplementing, and correcting OCR results from Azure Document Intelligence with Grooper's internal OCR engines (such as Transym and Tesseract). It provides configuration options, vocabulary support, and preferred pattern logic to improve the accuracy and consistency of extracted text.

Usage

• Used by AzureAIOcrChopper and related classes to align Azure OCR results with Grooper's OCR segments.
• Supports configuration of vocabulary, preferred patterns, and correction logic.
• Enables advanced diagnostics and annotation for review and troubleshooting.

Properties

Name	Type		Description
General
Traditional Ocr Profile	OCR Profile	►	An optional OCR Profile to be used when performing traditional OCR. EditorNode Reference Editor If not set, basic Transym OCR will be used with the default settings.
Vocabulary	Lexicon	►	An optional vocabulary lexicon defining words for the configured OCR language. EditorNode Reference Editor If set, this lexicon is used to break ties in cases where Azure and Transym disagree on the value of a text segment.
Options
Preferred Patterns	String	►	A regular expression defining one or more text patterns which should be preferred in cases where Azure and Transym disagree on the value of a text segment. EditorRegex Property Editor This regular expression should define patterns which occur frequently in the document set. For example, if processing invoices, currency values occur very frequently. The regular expression $?\d{1,3}(,?\d{3})*.\d{2} would ensure that '$5.00' is preferred over '$5.0o' or "$5,00", in the event that the OCR engines disagree. Use the regular expression 'OR' operator (|) to include multiple patterns.
Minimum Word Confidence	Double	►	The minimum confidence for text segments. ConverterPercent Converter Default0%
Minimum Segment Confidence	Double	►	The minimum confidence for a detected segment. ConverterPercent Converter Default0%
Suppress Handwriting	Boolean	►	If true, handwriting segments will be ignored. DefaultFalse
Ignore Checkboxes	Boolean	►	If enabled, detected checkboxes will be ignored. DefaultFalse Not all Azure models will detect checkboxes. Review the Microsoft documentation for more model specific details.

See Also

OCR ProfileLexicon

Used By

Azure DI OCR