MedImageOCR

Optical Character Recognition for Medical Images.

ocr.py

Uses Google's Tesseract engine for OCR.

Installation

• OpenCV ** pip install opencv-contrib-python for full install or pip install opencv-python for partial
• pytesseract ** pip install pytesseract

ocr_easy.py

Build image

docker build -t ocr:latest .

Build without a cache

docker build --no-cache -t ocr:latest .

Run

docker run --rm -v <path to data directory>:/data ocr:latest --zip /data/<zip file>

Run interactively

docker run --rm -v <path to data directory>:/data --env "PYTHONUNBUFFERED=1" -it --entrypoint /bin/bash ocr:latest

Run with GPU

docker run --rm --gpus all -v <path to data directory>:/data ocr:latest --zip /data/<zip file>