Chenyu

> Could anyone reproduce similar results on tanks&temples, e.g. meeting room and courtroom? I used numerical graident, coarse-to-fine optimization of grid and appearance embedding. however, i did not observe explicit...

Hi, @bennyguo Thanks for your reply! ``"... but I think it could be harder to converge than using volume to model the background"`` - That's exactly what confuses me a...

It is a valid point. Actually, I tried to model NeuS without nerf for backgrounds on unbounded scenes, but the reconstruction quality is really poor.

> > Hi, @AIBluefisher! In the NeuS system, the foreground is modeled as a signed distance field, but the background is often modeled as a volumetric part, so we need...

Theoretically speaking, it is unnecessary to model the background if we have the foreground/background masks. If we have a foreground point cloud, say it is dense enough, we can actually...

$VOC_TREE_PATH denotes the absolute path of the vocabulary tree provided by [COLMAP](https://demuc.de/colmap/). The second parameter denotes the upper bound of the images you want to match (you need to adjust...

you're right. Which file is it? I'll fix it soon.

It may be due to the view graph being too sparse and it contains multiple connected components. It can try to identify if there exists more than one connected component...

I'm not sure whether the current COLMAP provides relative poses. You may need to check with the relative poses of the official COLMAP.

This actually could happen when node indices are not consistent (something might be wrong with the input) or the view graph is not fully connected (has more than 1 connected...