I ask, because you then go on to sample from the matches based on that certainty which can result in points with certainty less than 0.05 being sampled. In fact, if all the matches have certainty below 0.05, then expansion_factor*num matches would still be selected.
Yes, this is intended. Although depending on the task you might want to use different types of thresholding. We found in general that it was better to "oversample" low confidence matches.
In DKM's sample method, do you intend to set any certainty greater than 0.05 to 1, instead of, perhaps, setting any certainty less than 0.05 to 0?
https://github.com/Parskatt/DKM/blob/fb85d632c5aaa5be643717701016f31d3ac0af22/dkm/models/dkm.py#L574
I ask, because you then go on to sample from the matches based on that certainty which can result in points with certainty less than 0.05 being sampled. In fact, if all the matches have certainty below 0.05, then
expansion_factor*num
matches would still be selected.