Quick preprocessing of springboard screenshots to make OCR easy. Basically, the only novelty here is checking (or more accurately, guessing) whether the text will be black or white. A mask is made to try to get rid of the background colors, and then OCRed and results tallied.

from springboard_ocr import SpringboardOcr
import json
so = SpringboardOcr()
so.process_image_file("test.png")
print(json.dumps(so.get_app_stats()))

returns

{
  "Mobile": 1,
  "BLAST": 1,
  "V'": 1,
  "Entertainment": 1,
  "Chrome": 1,
  "Duo": 1,
  "YouTube": 1,
  "Surfline": 1,
  "been": 1,
  "Netflix": 1,
  "h": 1,
  "News": 1,
  "NPR": 1,
  "OpenTable": 1,
  "OED": 1
}

• Install by doing a brew install tesseract, or follow this guide for directions on your system.

• Install python requirements by doing a pip install -r requirements.txt --user or however you wish manage your packages.