What is the purpose of the neural image function? #521
Unanswered
VisionaryMind
asked this question in
Q&A
Replies: 1 comment
-
Hi there, the image approximation task is about representing the image through a MLP—not about hallucinating extra detail. So the best case result is indeed that the neural image matches the target image almost exactly. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
NeRF works splendidly, even on very small datasets, but I am trying to figure out how to duplicate your results with neural images. It does not appear to be able to increase resolution on low-res images. The Einstein.exr example, for instance -- zooming in does nothing differently than if the image were loaded in a viewer and zoomed in on. Here is what I see:
I have read over your documentation several times, but cannot unravel the mystery of why neural imaging does not work. Even with the gigapixel Tokyo sample, resolution is progressively lost upon zoom-in.
Beta Was this translation helpful? Give feedback.
All reactions