sniklaus / softmax-splatting

an implementation of softmax splatting for differentiable forward warping using PyTorch
466 stars 58 forks source link

Question about the paragraph containing Eq. 13 #55

Closed seb-le closed 2 years ago

seb-le commented 2 years ago

First of all, thank you for your nice work.

I have a question about the paragraph (Softmax splatting. on page 4) containing Eq. 13 in the paper.

I think if we use depth as Z, the grass will appear instead of the car in Fig. 3 because distant objects (i.e., grass) are more important than the close object (i.e., car).

How can I understand this?

Thank you

sniklaus commented 2 years ago

Thank you for your interest in our paper! If you want to use depth, you can just provide -depth as the metric. However, depending on what you are working on you may want to use disparity instead (and disparity comes with the added bonus that you don't have to flip the sign).