/QTextRecognizer

A gui for tesseractOCR with some preprocessing image options (OpenCV) for improve character recognition

Primary LanguageCMIT LicenseMIT

QTextRecognizer

A gui for tesseractOCR with some preprocessing image options with OpenCV for better recognize text from images

GitHub top language GitHub All Releases GitHub Releases GitHub release (latest by date) GitHub Release Date

GitHub Releases

Additional downloads

  • You have to download languages from tesseract-ocr tessdata github page.

How to use?

  • Open the app, drag and drop picture or use open button.
  • Apply the preprocess options you want:
    • Rotate
    • Remove warp
    • Threshold
    • Brightness, contrast
  • Select the text blocks by mouse also press SHIFT key for select multiple text blocks.
  • Press Recognize text button and adjust the options you want:
    • Select language(s)
    • Select Page segmentation mode
    • Insert black/white list for characters you want
    • Select adaptive threshold before recognition you want
    • Press Start OCR and wait for the result

Screenshots

Main screen

Rotate image screen

Deskew image screen

Image adjustments screen

Autodetect text layout

Recognize text screen

Settings screen and dark skin

Building

Dependencies

  • Qt 5.x and Qt Creator
  • OpenCV 4.x
  • Tesseract 4.1
  • Leptonica