You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Very wonderful work!
I want to know about the use of sky_seg=args.sky_seg, load_normal=args.load_normal, load_depth=args.load_depth. When I only have colmap sparsely reconstructed data, these parts will not be used, so what is the impact on the results?
In addition, the paper mentioned that Progressive Gaussian Propagation is very beneficial to the geometric consistency of low-texture areas. I would like to know if you have tried it on an indoor data set, which contains a large number of corridor areas with no texture or weak texture.
Very much looking forward to your reply
The text was updated successfully, but these errors were encountered:
load_normal=args.load_normal, load_depth=args.load_depth is designed for some ablation using the monocular priors. It has no affect if you do not use them. If you use the outdoor scene, it is better to get the sky mask first.
It could be helpful if using the indoor scene. Additionally, it will be much more better if enhancing our propagation with a plane-prior patch-matching method.
Very wonderful work!
I want to know about the use of sky_seg=args.sky_seg, load_normal=args.load_normal, load_depth=args.load_depth. When I only have colmap sparsely reconstructed data, these parts will not be used, so what is the impact on the results?
In addition, the paper mentioned that Progressive Gaussian Propagation is very beneficial to the geometric consistency of low-texture areas. I would like to know if you have tried it on an indoor data set, which contains a large number of corridor areas with no texture or weak texture.
Very much looking forward to your reply
The text was updated successfully, but these errors were encountered: