Link Search Menu Expand Document

IBaseOCRExtractor Interface

Defines a base interface for PDF text extractors.

Assembly: Bytescout.PDFExtractor (in Bytescout.PDFExtractor.dll) Version:
public interface IBaseOCRExtractor

The IBaseOCRExtractor type exposes the following members.

Public propertyOCRAutoModeMinExistingTextLength
Applies to OCRMode.Auto only. Sets the minimal total length of normal text objects on the current page to decide whether to perform OCR or not. If the length of the text is less than this value then extractor will use OCR for this page. Otherwise it extracts the existing normal text objects only.
Public propertyOCRBlackList
A set of characters not allowed to be recognized from scanned document. The resulting text will only contain characters that are not in this list. This helps improve uncertain recognition.
Public propertyOCRCacheMode
Sets caching of OCR results. Default is Off (no caching). In WholePage caching mode you can save processing time as the SDK will check if need to re-run OCR on the page or it can just pull results from previously cached OCR results.
Public propertyOCRCorrections
A set of corrections automatically applied to recognized text to fix repeating recognition errors.
Public propertyOCRDetectLines
Sets whether to detect horizontal and vertical lines during OCR. They help to detect the table structure in scanned documents. Default is false.
Public propertyOCRDetectPageRotation
Sets whether to detect scanned page rotation. Slightly slows down the processing.
Public propertyOCRDisableAutoSegmentation
Sets whether to disable the automatic segmentation. All pages will be treated as containing sparse text (as a typical document such as an invoice).
Public propertyOCRImagePreprocessingFilters
A set of corrections automatically applied to recognized text to fix repeating recognition errors.
Public propertyOCRLanguage
Language of OCR (Optical Character Recognition). The valid values are:
  • "eng" - English (default)
  • "deu" - German
  • "fra" - French
  • "spa" - Spanish

Download more languages at

Public propertyOCRLanguageDataFolder
Folder containing OCR language data files.
Public propertyOCRMaximizeCPUUtilization
Gets or sets maximum OCR performance using Intel OpenMP (if available) to accelerate to approximately 30%. Default is false.
Public propertyOCRMode
Enables use of the OCR (Optical Character Recognition) which allows you to recognize text from images embedded in PDF pages. Set the language using the OCRLanguage property.

This option requires Bytescout.PDFExtractor.OCRExtension.dll assembly reference and relevant language files in OCRLanguageDataFolder folder. The SDK is shipped with language files for English, French, German and Spanish. You can download more languages at

Public propertyOCROverallConfidence
Gets overall OCR confidence. This value is calculated after the extraction is performed.
Public propertyOCRPageCount
Gets count of processed pages for which OCR was performed. This value is calculated after the extraction is finished.
Public propertyOCRResolution
Resolution of OCR (Optical Character Recognition). Default is 300 DPI.
Public propertyOCRWhiteList
A set of characters allowed to be recognized from scanned document. Only characters from this list will appear in the result text. This helps improve uncertain recognition.
Public methodIsOCRRecommendedForPage
Detects whether OCR is recommended for specified page. OCR (Optical Character Recognition) is recommended when pages has no text objects bat has an image that might contain text.
See Also