To detect documents in trivial images, process them and extract text from them. Used OpenCV v3.0 for document processing and Tesseract v3.03 for optical text recognition

Usage: python docscan.py -i data/Sample1.jpg

Roadmap:

• Release standalone Linux desktop version and smartphone versions of this project.
• Improve the accuracy of the project.
  • Possible areas for imrpovement:
    • Automate some parts of Tesseract training module for easier training.
    • Use different levels of smoothing on Tesseract input, compare text outputs and merge them so as to get high accuracy.
    • Extensively learn Tesseract's methods to try and modify them for greater accuracy.