Skip to content

README.rst

Latest commit

 

History

History
60 lines (43 loc) · 2.35 KB

README.rst

File metadata and controls

60 lines (43 loc) · 2.35 KB

pythOCR

Contents

Overview

This project intends to accomplish the following goals

  • Ability to recognize handwritten numbers and arithmetic operators (+, -, /, *, =) in real-time from a video feed
  • Arithmetic expressions are parsed by the software into a plain-text format
  • The expression is sent to WolframAlpha (api?)
  • The result is displayed on the screen

Roadmap

  • Ability to read expressions from video feed
    • efficient means to extract video frames and resample if needed
    • fast edge detection on said frames
    • custom algorithm needed for symbol detection
      • Standard OCR will not work (eg. pyTesseract)
    • validation & sanity checking
  • Ability to evaluate expression through Wolfram (or similar)
    • API available?
    • If not, is the webpage scrape-able
  • Ability to display results
    • Load browser in frame with query result?
      • Possibly easiest method; query can be inserted into URL without needing any API call
    • Parse results and convert to alternate format?

Tooling

Tool Function
Python 3.X implementation language
scikit edge detection
numpy general purpose math library
umlet diagramming
reStructuredText documentation language