Feature enhancement

[rajivupreti]

Thank you for providing great library.
Apologies for lengthy question.
It will be highly appreciated if you could provide your opinions on my following novice queries.

I am doing PoC of an application which needs to scan documents of varying size.
In my use case, I need to scan smaller size to larger size (sometimes size size is larger than A4).

While doing PoC, I noticed following.

1. If document is bigger in dimension (e.g. width: 5cm, height:37cm), edge was not detected properly
   and detected edge falls back to whole screen.

2. Edge detection does not work properly when the background is not dark enough.

Currently, I am trying to figure out whether above enhancement is feasible or not.
It will be highly appreciated if you could share opinion/hints on my following query.

• Currently it uses OpenCV Canny Detection if I am not wrong.

  Is it possible to switch to another Contour detection library?
  What sorts of changes are required in that case or any recommended library to use.

• To make the edge detection support for larger size document, what can be the enhancements to be considered.
  Do we need to think on using other Edge detection technology like Yolo etc. Or, can it be achieved using OpenCV?