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
I’m encountering an issue where the algorithm does not converge when using sparse view inputs (e.g., 6 or even 27 images). It works fine with dense inputs but fails with sparse ones.
Questions:
Are there specific settings needed for sparse views?
Are there any known issues with handling sparse inputs?
Best regards,
Hwan
The text was updated successfully, but these errors were encountered:
thanks for raising this topic. I think that behaviour is somewhat expected. Since this might be interesting for a wider audience, I added the following text to our FAQ section in this repo's README.
Q: Is ACE0 able to reconstruct from a small set of sparse views?
A: It can work but this scenario is challenging for ACE0. We expect other methods, and even COLMAP, to work much better in this case. ACE0 relies on images having sufficient visual overlap, particularly when registering new images to the reconstruction. You can lower the registration threshold when running ace_zero.py via --registration_confidence setting it to 300 or 100 - but at some point ACE0 will get unstable. ACE0 shines if you have dense coverage of a scene, and reconstruct it from many images in reasonable time.
Hi,
Thanks for the great work on ACE0!
I’m encountering an issue where the algorithm does not converge when using sparse view inputs (e.g., 6 or even 27 images). It works fine with dense inputs but fails with sparse ones.
Questions:
Best regards,
Hwan
The text was updated successfully, but these errors were encountered: