I have a code base using OpenCV to basically perform a tiny scanner like operation. Take a picture, detect the foreground document and warp the object to shape and applying the Black and White effect to it.
That said, the detection of the foreground doc can be improved a lot. We will be giving you a series of documents under various lightly conditions and your job is to optimize the code to capture most of the scenarios.
It should be super quick for those who have done it before or have some OpenCV experience.
This is mainly for optimization of existing python codes. PM me for the codes.