training with vlp-16 dataset about self-supervised-depth-completion HOT 7 OPEN

For cameras with different intrinsic parameters than the KITTI cameras, there is a slightly more complicated image resizing process than simple cropping. In particular, we need to resize the image in a way such that the resized image has the same effective intrinsics as KITTI.

from self-supervised-depth-completion.

Hi Fangchang,
I have the same question. Why can't we just use our own intrinsic parameters instead of resizing images? Is it related to some hyperparameters needed to tune?

from self-supervised-depth-completion.

@XiaotaoGuo It might or might not work well with a different set of intrinsics, but there is simply no guarantee that the trained network would transfer directly to this new set of intrinsics (and image sizes). My suggestion is to keep the test images as close to the train images as possible.

from self-supervised-depth-completion.

Thanks! What if we use our own dataset to train the network and test with it?

from self-supervised-depth-completion.

What if we use our own dataset to train the network and test with it?

Then there is no need for any resizing

from self-supervised-depth-completion.

Hi @Melvintt,
Did you manage to work this out? I am fighting similar problems with my own dataset, generated with a VLP32 and a Zed camera.

from self-supervised-depth-completion.

What if we use our own dataset to train the network and test with it?

Then there is no need for any resizing

Hi，
When we use 'resize' operation, how to deal with the sparse depth?
Could it be correct to just resize the sparse depth input like the rgb images, although it is sparse?
Thanks!

from self-supervised-depth-completion.