Calibration Free Imaging

[dennisgorelik]

It gets better:
~~~~~~~~~~~
https://youtu.be/UGL_OL3OrCE?t=1177
19:37
What you can do is to use methods where you [have] do not need any calibration whatsoever and you can still can get pretty good results.
So here on the bottom at the top is the truth image, and this is simulated data, as we are increasing the amount of amplitude error and you can see here ... it's hard to see ... but it breaks down once you add too much gain here. But if we use just closure quantities - we are invariant to that.
So that really, actually, been a really huge step for the project, because we had such bad gains.
~~~~~~~~~~~

"Вот тут мне карта и поперла".


==============
https://youtu.be/UGL_OL3OrCE?t=2242
37:22
And you can notice like at the bottom we get really terrible reconstruction, just cause if it fits the data very well, because you know it maybe wants to smooth out the flux as much as possible and we don't select things like that in the true data.
==============

I posted a comment below that video:
~~~~~~~~~
19:39 "Calibration Free Imaging"
Does it mean that you were using measurements tools (telescopes) without prior calibration?
~~~~~~~~~
but it got deleted... twice.

Update:
Discussing "not need any calibration whatsoever" on HN
"Calibration Free Imaging" (AKA scam)

Comments

[symbioid]

I found the paper if you are curious about the algorithm itself:

http://people.csail.mit.edu/klbouman/pw/papers_and_presentations/cvpr2016_bouman.pdf

[symbioid]

Obviously the algorithm isn't tied to the the process of this particular experiment, so it seems unrelated to your complaints here, but still figured it's worth sharing.

[dennisgorelik]

Did you personally learn anything interesting from that paper?

[dennisgorelik]

I saw this paper already, but it is not clear what I can do about it.

The paper is a bit old (2016) and outlines their plans of how to extract image.

It shows that from the beginning they were determined to introduce "prior" bias into their image interpretation:
~~~~~~~~~~~~~~~~~
http://people.csail.mit.edu/klbouman/pw/papers_and_presentations/cvpr2016_bouman.pdf
4. Method
Reconstructing an image using bispectrum measurements is an ill-posed problem, and as such there are an infinite number of possible images that explain the data [28].
The challenge is to find an explanation that respects our
prior assumptions about the “visual” universe while still satisfying the observed data.
~~~~~~~~~~~~~~~~~