Image to Text Extractor

An OCR (Optical Character Recognition) application that extracts text from images using Tesseract OCR engine with a user-friendly GUI.

Author

Paradorn Katananon

Features

• 📷 Single Image Processing: Extract text from individual images
• 📁 Batch Processing: Process multiple images in a folder at once
• ✏️ Editable Preview: Review and edit extracted text before saving
• 🌍 Multi-language Support: English, Chinese (Simplified/Traditional), Spanish, French, German, Japanese, Korean
• ⚙️ Configurable OCR Settings: Adjust page segmentation modes for better accuracy
• 💾 Flexible Saving: Save individual files or batch process with automatic file naming
• 🔄 Responsive UI: Multi-threaded processing keeps the interface responsive

Requirements

• Python 3.x
• Tesseract OCR engine
• Required Python packages:
  • pytesseract
  • Pillow (PIL)
  • tkinter (usually included with Python)

Installation

1. Install Tesseract OCR:

   • Windows: Download from GitHub Tesseract Releases
   • macOS: brew install tesseract
   • Linux: sudo apt-get install tesseract-ocr

2. Install Python dependencies:

   pip install pytesseract Pillow

3. Configure Tesseract path (if needed): If Tesseract is not in your system PATH, add this line to the code:

   pytesseract.pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'

Usage

Running the Application

python img2txt.py

Single Image Mode

1. Click "📷 Select Single Image"
2. Choose an image file (PNG, JPG, JPEG, TIFF, BMP, GIF)
3. Review the extracted text in the preview area
4. Edit the text if needed
5. Click "💾 Save Text to File" to save

Batch Processing Mode

1. Click "📁 Select Folder (Batch)"
2. Select a folder containing multiple images
3. All images will be processed automatically
4. Text files are saved in an extracted_texts subfolder
5. Review the processing summary

OCR Settings

Language Selection

Choose the appropriate language for better OCR accuracy:

• eng - English
• chi_sim - Chinese Simplified
• chi_tra - Chinese Traditional
• spa - Spanish
• fra - French
• deu - German
• jpn - Japanese
• kor - Korean

Page Segmentation Modes

Select the page segmentation mode based on your image type:

• PSM 3 (Auto): Fully automatic page segmentation. Best for general documents when you're unsure.
• PSM 6 (Block): Single uniform block of text. Ideal for clean documents, books, single-column text.
• PSM 11 (Sparse): Sparse text with no particular order. Good for screenshots, forms, receipts, or scattered text.
• PSM 12 (Sparse + OSD): Sparse text with orientation and script detection. Same as PSM 11 but handles rotated text.

Supported Image Formats

• PNG
• JPG/JPEG
• TIFF
• BMP
• GIF

File Structure

img2txt/
├── img2txt.py          # Main application file
├── README.md           # This file
└── extracted_texts/    # Created automatically for batch processing

Tips for Better OCR Results

1. Use high-resolution images: Better image quality = better text recognition
2. Ensure good contrast: Black text on white background works best
3. Avoid skewed images: Straighten images for better accuracy
4. Choose correct language: Select the language that matches your image text
5. Adjust PSM mode: Experiment with different segmentation modes for your specific use case

Troubleshooting

"Tesseract not found" error

• Ensure Tesseract OCR is installed
• Add Tesseract to your system PATH
• Or specify the path in the code

Poor OCR accuracy

• Check image quality and resolution
• Try different page segmentation modes
• Ensure correct language is selected
• Pre-process images (enhance contrast, remove noise)

Language not working

• Install additional Tesseract language packs
• Verify language data files are in Tesseract's tessdata folder

License

This project is licensed under the MIT License. See the LICENSE file for details.

Version

1.0

---

Created by Paradorn Katananon