Hi, thanks for the great work!
I am getting reasonable results on my own data, but I am finding that bad depth maps are causing artifacts in the final renders.
For example:
What is the best way to get good depth maps with the pre-trained TransMVSNet depth model?
Does it work best at a certain input resolution?
Hi, thanks for the great work! I am getting reasonable results on my own data, but I am finding that bad depth maps are causing artifacts in the final renders. For example:
What is the best way to get good depth maps with the pre-trained TransMVSNet depth model? Does it work best at a certain input resolution?