Open eezywu opened 5 years ago
Hi, I was confused that why a single conv layer is able to learn the offsets, since a conv layer is local connected to the input. So how could it learn the offsets that point to the other positions? Thanks!
Can anyone explain? Thanks!
Hi, I have the same question. Have you figured it out?
can we use deform conv to generate offsets?
Hi, I was confused that why a single conv layer is able to learn the offsets, since a conv layer is local connected to the input. So how could it learn the offsets that point to the other positions? Thanks!