princeton-vl/RAFT-Stereo

Why divide the correlation by sqrt(D)?

steven9046 opened this issue · 1 comments

steven9046 commented

RAFT-Stereo/core/corr.py

Line 156 in 0e2a127

return corr / torch.sqrt(torch.tensor(D).float())

Hi, what does this line mean? In the paper it says that the correlation is dot product between feature vectors. But it is divided by this sqrt. Any meaning? Can replace this sqrt by something else?

lahavlipson commented

This reduces the variance in the correlation dot products. But it is not necessary and if you remove it and re-train the model, the performance should be similar.