Skip to main content

Tesseract Font Editor Usage

tip
  • Trains Tesseract font libraries; based on jTessBoxEditor with simplified workflow for easier editing
  • Trained Tesseract libraries work with EC Android, iOS USB, iOS standalone, HarmonyOS Next, etc.

Download & Installโ€‹

  • EC cloud drive โ†’ Dev Tools โ†’ OCR Resources โ†’ download EasyClick-TesseractOCR-ๅญ—ๅบ“็ผ–่พ‘ๅ™จ.zip
  • Extract and run tessocr-editor.exe
  • Note: extract path must not contain Chinese, spaces, or special characters โ€” otherwise validation may fail
  • Download area: Download Area

Train Font Libraryโ€‹

  • Open tessocr-editor.exe โ€” interface:
  • Tesseract path โ€” leave default (bundled tesseract). If changing, point to tesseract.exe

Step 1: Merge Images to TIF Buttonโ€‹

  • Merges jpg, png, bmp into a tif file โ€” select folder or single image
  • Example: folder E:/jtess/img/, new font name newfont โ€” log shows ็”Ÿๆˆtifๆ–‡ไปถๆˆๅŠŸ (TIF generated successfully)

Step 2: Select TIF Fileโ€‹

  • Click select button to choose tif file

Step 3: Generate box Font Fileโ€‹

  • Select Generate box file, click Start. Default PSM 6, base language chi_sim+eng (Chinese simplified + English)
  • box generation log
  • System creates a box file in the tif folder

Step 4: Train Font Libraryโ€‹

  • Generates Tesseract traineddata file
  • Select Generate traineddata training set file, click Start; log shows traineddata output path

Step 5: Validateโ€‹

  • Click Validate, pick an image to test the trained library
  • Can clear Base language set value for validation
  • If results are poor, use Font Editor tab to edit box and retrain/validate

Edit Font Libraryโ€‹

  • Interface:

Open box Fileโ€‹

  • Click Open (shortcut ctrl+o), select a tif file โ€” box opens automatically
  • Left table: characters and coordinates; right: image and character boxes

Edit Textโ€‹

  • Double-click table text and press Enter to edit
  • Or edit Character field and coordinates at top

Change Character Box Coordinatesโ€‹

  • Select table row โ€” character highlighted with red box on image
  • Drag red box corner arrows to resize
  • Drag red box to move
  • Or use arrow keys to adjust coordinates

Zoomed Character Viewโ€‹

  • Click Character View tab; arrow keys move red box
  • ctrl+arrow zoom box in/out until text is fully enclosed

Image Editingโ€‹

  • Uses OpenCV for grayscale and binarization when training/recognition quality is poor

Other Featuresโ€‹

  • Merge: merge multiple box characters
  • Split: split one box into two
  • Insert: new box
  • Delete: remove box
  • Font: table font setting (optional)
  • box table right-click menu has additional actions

Shortcutsโ€‹

  • ctrl+o โ€” open tif/box
  • ctrl+s โ€” save box
  • ctrl+p or alt+p โ€” previous character
  • ctrl+n or alt+n โ€” next character
  • Arrow keys โ€” move box
  • ctrl+arrow or alt+arrow โ€” resize box
  • Mouse drag โ€” move box
  • Mouse drag corner โ€” resize box
  • ctrl+d or alt+d โ€” previous image
  • ctrl+g or alt+g โ€” next image

Usageโ€‹

  • After editing and retraining to traineddata, use per tessocr tutorials

Saveโ€‹

  • Save box edits before closing to avoid data loss

FAQโ€‹

  • Validation fails
  • Extract path may contain Chinese/special chars โ€” use English path
  • App crashes on open
  • Install VC runtime from cloud drive, restart PC
  • Training error FAILURE! couldn't find a matching blob
  • Tesseract cannot match shapes/features in the image. Try:
  • Higher image quality โ€” clear, good contrast
  • Preprocessing โ€” binarization, denoise, deskew
  • Tune Tesseract params โ€” psm, oem
  • Adjust box ranges