It's a very interesting work! But I have a question about the coarse depth map sampling. In this part of the code, why is 15 used as the threshold? What factors are considered in this decision?
Hi,
15 is implemented by https://github.com/JonathonLuiten/diff-gaussian-rasterization-w-depth
it means there are holes/invalid pixels in the depth map. so we generate a slectmask for valid depth only.
It's a very interesting work! But I have a question about the coarse depth map sampling. In this part of the code, why is 15 used as the threshold? What factors are considered in this decision?