Document Boundary & Canny Edge Detection using OpenCV
- Used Simple OpenCV Library for Boundary and Edge Detection
- The Algorithm is quite fast and does not require any training and works quite well with rectangle documents as it was meant for document boundry detection
- Did not use Deep Learning Techniques as it was most computationally expensive and required good Dataset of images with Boundry label and would had been slow too
- Sometimes simpler models are better than deep learning models
Make Sure you have following libraries installed
- opencv
- imutils
Use pip install -r requirements.txt to get all the dependencies
To Run
Use python scan.py --image images/1.jpg to train the model
It opens two windows one with edge detection and one with boundary detection
(1) https://www.pyimagesearch.com/2014/09/01/build-kick-ass-mobile-document-scanner-just-5-minutes/