Recognition text on video

Plan:
1. Read frames from video
2. Preproc frames to lower resolution, change size and etc.
3. Recognize text from frames
4. Deduplicate text, proc, (optional check errors, try to fix them)
5. Save result

Requirements:
  Python on unix system:
    python3 -m venv .venv
    source .venv/bin/activate
    pip install -r requirements.txt
  System:
    MacOs:
      brew install tesseract
      brew install tesseract-lang
    Linux:
      sudo apt-get install tesseract-ocr