Startools V1.8 Spatially Variant PSF Deconvolution

User images created with StarTools.
User avatar
admin
Site Admin
Posts: 3381
Joined: Thu Dec 02, 2010 10:51 pm
Location: Melbourne
Contact:

Re: Startools V1.8 Spatially Variant PSF Deconvolution

Post by admin »

The reversal indeed pertains to reversing deconvolution (the reverse being convolution). See also the convolution theorem. E.g. the ability to "re-blur" the dataset using the used PSF and arrive at the original input dataset. It is also trivial to recover the exact PSF if input and output datasets are available.

Mathematically, this is a very simple affair, though as (most of you know), deconvolution in the real world is much harder without the solution destabilising due to the signal being imperfect. This does necessitate some intervention in the otherwise very basic process, hence iterative algorithms like Richardson & Lucy decon.

The important thing here is that you can articulate what is happening to the signal and "how you got there" - a prerequisite in science so the result can be scrutinized and replicated.

In the case of real deconvolution;
  • you are responsible for (and in control of) providing the PSF(s)
  • the algorithm then reverse-applies this/these PSFs in a well understood, well documented, and well-accepted manner
  • no new information is introduced, and all information comes from the observation/dataset itself
In the case of a neural net that is used to hallucinate an output in response to an input;
  • deconvolution is not explicitly implemented and there is no concept of a method, algorithm, let alone PSF that can be articulated, provided or extracted
  • the neural net re-interprets the image in a black box manner, without being able to articulate how and why some pixels were changed or how they relate to the input data
  • new information is introduced, and the information does not come from the observation/dataset itself
In the case of deconvolution, the result is a solution for a restoration that can be robustly defended with all the relevant information in hand. As such - if up to snuff - it retains or enhances documentary and scientific value.

In the case of neural net hallucination, the result is a re-interpretation of the dataset, that may or may not look like a restoration, that has no intrinsic evidence supporting its validity, method or origins. As such it destroys documentary and scientific value.
Ivo Jager
StarTools creator and astronomy enthusiast
Startrek
Posts: 401
Joined: Mon Dec 30, 2019 3:49 am

Re: Startools V1.8 Spatially Variant PSF Deconvolution

Post by Startrek »

Ivo,
An excellent descriptive evaluation of Deconvolution and Non Deconvolution processes
I posted my images on our local forum and indeed turned some heads around to take another look , mostly Pi and PS. users. One user agreed that from looking at my images, ST SV Decon does a much better job than BlurXterminator. He has used BlurXterminator quite extensively on his recent data sets
It Startools keeps improving with each version , traditional software apps may no longer be the popular norm in the future.

Clear Skies
Martin
Post Reply