TsgcAIOCRTesseract

Local OCR based on the Tesseract command line tool.

Introduction

TsgcAIOCRTesseract recognises the text of an image with the Tesseract OCR command line tool. It is the local OCR used by TsgcAIDocumentExtractor when OCRMode is domFallback or domAlways, and it can also be used on its own.

Requirements

Tesseract must be installed on the machine, with the trained data of every language you use. The component runs the tesseract executable and reads its standard output; nothing is linked into your application.
Images only: PNG, JPEG, TIFF, BMP and the other formats your Tesseract build reads. A PDF raises EsgcAIOCR; convert its pages to images first.
Windows, Linux and macOS. Android and iOS are not supported. On macOS the Delphi and C++Builder component does not apply TimeoutMs, the call waits until Tesseract ends. The .NET component applies it on every OS.

Properties

Property	Type	Default	Description
ExecutablePath	string	tesseract	Path of the Tesseract executable. The default finds it through the PATH.
Language	string	eng	Tesseract language codes, joined with + for several languages (for example `eng+spa`). Passed as `-l`.
PageSegMode	Integer	3	Tesseract page segmentation mode, passed as `--psm`. 3 is fully automatic, 6 treats the image as one block of text.
TimeoutMs	Integer	60000	Maximum time for one recognition, in milliseconds. 0 waits without limit. Not applied on macOS in Delphi and C++Builder; .NET applies it on every OS.
ExtraArguments	string		Extra command line arguments appended as they are, for example `--oem 1`.

Methods

Method	Description
`function RecognizeFile(const aFileName: string): string;`	Returns the text of the image. Raises EsgcAIOCR when the file does not exist, when it is a PDF, or when Tesseract ends with an error or does not end within TimeoutMs.
`function RecognizeStream(const aStream: TStream; const aExtension: string): string;`	Writes the stream to a temporary file with the given extension, recognises it and deletes the file.
`function IsAvailable: Boolean;`	Runs `tesseract --version` and returns True when it succeeds. Use it to check the installation before the first extraction.

Events

Event	Description
`OnOCRText(Sender: TObject; const aFileName: string; var aText: string)`	Raised with the recognised text before it is returned, so it can be cleaned or corrected.

Example

oOCR := TsgcAIOCRTesseract.Create(nil);
oOCR.Language := 'eng+spa';
if not oOCR.IsAvailable then
  raise Exception.Create('Tesseract is not installed');
ShowMessage(oOCR.RecognizeFile('receipt.png'));

oExtractor.OCR := oOCR;
oExtractor.OCRMode := domFallback;

TsgcAIOCRTesseract is available in Delphi, C++Builder and .NET. In .NET the methods are RecognizeFile(string aFileName), RecognizeStream(Stream aStream, string aExtension) and IsAvailable(), and the OnOCRText event passes the text as ref string aText.