OCR Tool

OCR application for text recognition from images using Tesseract.

Features

• Text recognition from images (PNG, JPG, JPEG, TIFF, BMP)
• Text recognition from PDF files
• Web interface for file upload
• Language selection for recognition (English, Russian)
• Export result (copy, download)

Docker

docker build -t ocr-tool .
docker run -p 8080:8080 ocr-tool

Open http://localhost:8080

Or local run (needs java 17 & maven installed)

mvn exec:java -D"exec.mainClass=com.github.dkrut.WebServer"

Web Interface Usage

1. Select file (drag & drop or "Select File" button)
2. Select language(s) - English, Russian
3. Click "Start OCR"
4. Copy or download result

Architecture

Classes

• Ocr - image processing via Tesseract (grayscale preprocessing)
• PdfConverter - PDF to PNG conversion
• WebServer - web interface (Javalin)

Processing Flow

1. File uploaded via web interface
2. Temporary folder created: TEMP/ocr-{uuid}/
3. For PDF: converted to PNG (pdf-images/)
4. Image converted to grayscale (grayscale/)
5. Text recognized by Tesseract
6. Temporary files deleted