Spell Corrector

Can be one of the following types:

The CorrectionExtractor is a Value Extractor that identifies and corrects spelling errors in OCR data.

• This extractor should be configured to match incorrect or misspelled values in the field's text and output the corrected value.
• During spell correction, each match found by the extractor is replaced with its corrected value.
• Character casing is preserved or adjusted automatically to match the original text where possible.

Example:

• Use a Pattern Match extractor to find common OCR errors such as "lnvoice" (with a lowercase 'L' instead of 'I') and output "Invoice".
• Use a List Match extractor to map a list of known misspellings to their correct forms.

If this property is not set, only general OCR repair options (such as and ) will be applied.

When enabled, this option attempts to correct case errors in OCR data, such as converting "iNVOICE" to "INVOICE" or "invoice" as appropriate.

• Use this when the OCR engine has difficulty distinguishing between upper and lower case letters, especially in all-caps or mixed-case documents.
• The repair logic analyzes the context and font metrics to determine the intended case for each character.

Example: > If the OCR output is "iNVOICE TOTAL", enabling 'RepairCase' will correct it to "INVOICE TOTAL".

When enabled, this option attempts to resolve ambiguous or commonly misread characters in OCR data, such as:

• 'O' (letter O) vs. '0' (zero)

• 'I' (uppercase i), 'l' (lowercase L), and '1' (one)

• 'S' (letter S) vs. '5' (five)

• Use this when the OCR engine frequently confuses these characters, especially in numeric fields, codes, or identifiers.

• The repair logic uses character context and patterns to make corrections.

Example: > If the OCR output is "S1O5", enabling 'Disambiguate' may correct it to "5105" if the context suggests a numeric value.