Running Euclid no-detection fitting test data set

[ramakr18]

Hello,

I'm trying to run the Euclid Q1 model fitting demo for the Sersic paramters, and I am seeing that the results of the demo have a surprisingly poor agreement with the values made available on the ESA website. I cross-matched the output catalog shared as part of the demo against the ESA catalog, and found that many of the sources show a large disagreement in the Sersic parameters between the two catalogs. I am confused by this - I was under the impression that the demo provided the exact settings used to do the Euclid Q1 model fitting, and so the results should be exactly the same, or at least very similar. Could you tell me what the possible sources of disagreement are?

[mkuemmel]

Hi @ramakr18,

nice catch. Please note that the README says:

The PSF models used in this test dataset are slightly simplified compared to the Q1 processing. While the results remain reasonable, they are not identical to those from Q1.

Later I realized that the (single) PSF I chose for this set is actually pretty, as can be seen on the VIS PSF:

That truncated PSF (it's the same in YJH) can not yield reasonable result and consequently the results differ from Q1.

The setting is identical to Q1, but the input data is not.

Hope this explains the situation.

[mkuemmel]

@ramakr18 can we close this??