Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Vkitti2 - sky part wrong depth #220

Open
hgolestaniii opened this issue Dec 4, 2024 · 2 comments
Open

Vkitti2 - sky part wrong depth #220

hgolestaniii opened this issue Dec 4, 2024 · 2 comments

Comments

@hgolestaniii
Copy link

Hi,

I tried depth anything v2 on vKitti2 in two different modes:
1- the general mode which outputs relative depth data
2- metric depth estimation fine-tuned on vKitti2

I got very strange depth for sky region for the "metric" one. Please see the attached image. Why is it the case? The network is fine-tuned on vKitti2, right? Then it should give us very good estimation for this dataset; however, the sky is quite off.

Q1- Have you excluded the sky region (or anything further than 80m) during fine-tuning?
Q2- Is it also the case for Depth AnyThing v1?

skyIssue_vKitti2
@hcscysh
Copy link

hcscysh commented Dec 6, 2024

I want to know whether the fine-tuning network you mentioned means retraining under the weights loaded in the demo, or freezing some network parameters

@LiheYoung
Copy link
Contributor

Hi @hgolestaniii, our metric depth model trained on VKITTI 2 can only estimate depth ranging from 0m to 80m, since we clip large depth values (> 80m) during training.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants