IronOcrEngineParameters

AcceptedOcrCharacters

Specifies the characters the OCR engine will look for, which can hugely improve speed and accuracy. The default behavior is to identify all UTF-8 characters including accented characters borrowed from international languages. E.g. "0123456789" for only numbers, or "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz.:;?!@$%&+-,-()\"'0123456789" for English characters without accents. Default: null

Syntax: public string AcceptedOcrCharacters { get; set; }

CleanBackgroundNoise

AdvancedOcr will clean background 'noise' , dust, speckles and signs of paper away from printed text. In the context of detecting text on the UI, CleanBackgroundNoise can help on screen segments with different backgrounds, e.g. tables with alternating row colors. Note: This will produce unusable results for white text on dark backgrounds. If you want to use it in that case, enable UseInvertModification in PreprocessingParameters. Default: false

Syntax: public bool CleanBackgroundNoise { get; set; }

ColorDepth

Bits of color per channel per pixel. Lower ColorDepth can reduce scanning time for documents that also contain many colors or images. A value of ZERO will result in image color depths not being modified from the original source image. Default: 0

Syntax: public int ColorDepth { get; set; }

ColorSpace

OCR documents can be scanned in GreyScale to reduce processing time. This may not work well for low-contrast text on a non-white backgrounds. Default: GrayScale

Syntax: public OcrColorSpace ColorSpace { get; set; }

DetectWhiteTextOnDarkBackgrounds

Automatically detects and adjusts white text on darker backgrounds. Default: true

Syntax: public bool DetectWhiteTextOnDarkBackgrounds { get; set; }

EnhanceContrast

Enhances image contrast adaptively, which generally increases text readability both by humans and machines. Default: false

Syntax: public bool EnhanceContrast { get; set; }

EnhanceResolution

Re-renders low quality images at a higher DPI. IronOcr works faster and more accurately with high resolution images. Useful for images scanned at under 250 DPI (standard dpi for vnc is 95, so this is a very important parameter for detecting text on the UI!) Default: true

Syntax: public bool EnhanceResolution { get; set; }

InputImageType

Hints the OCR Engine about the type of image being scanned, which can reduce OCR time and increase accuracy. Can be AutoDetect, Document, Snippet Default: Snippet

Syntax: public InputTypes InputImageType { get; set; }

ReadBarCodes

Enables Barcodes to be extracted from OCR documents, alongside the text content. Default: false

Syntax: public bool ReadBarCodes { get; set; }

RotateAndStraighten

AdvancedOcr will deskew and rotate images to be straight, upright, unbent and without perspective. RotateAndStraighten is a powerful tool, making imperfect scans OCR accurately. CPU/GPU intensive. Default: false

Syntax: public bool RotateAndStraighten { get; set; }

Strategy

Algorithm to be used for OCR text analysis. Can be Fast or Advanced. Advanced will take considerably longer and produce much more false results, but might help to recognize text that is not detected as text with Fast strategy. Default: Fast

Syntax: public OcrStrategy Strategy { get; set; }

UseAutoMode

AutoOcr uses meta-data and computer-vision to predict intelligent default settings for each setting. Ignores all other IronOcrParameters except ReadBarCodes. Generally takes longer than the default settings for AdvancedOcr and doesn't produce better results. Default: false

Syntax: public bool UseAutoMode { get; set; }