Hello,
While going through the recording, in the Hough transform explanation part, I noticed that we do not necessarily care about the value of the gradient itself (when getting the edges before voting). As was mentioned, we do some thresholding on the gradients to avoid irrelevant lines/shapes, but would there be a case where weighting/giving more importance (when voting) to some points because of a very large gradients on those points is beneficial/needed, and if so how exactly would we proceed.
For the large gradients set of points case, that I'm mentioning, I first thought about delineating a small rectangular box (very contrasted) by simply using the linear hough transform (which we could not delimitate into segments, because the edges would not even make it to the top k, perhaps general hough transform could solve this issue of the box, but that's why I thought, perhaps there is an importance to the value of the gradient).
Thanks you in advance for the answer
