StarTools

Hi all,

I repeat the same question I did in another forum.
I don't know still the difference between binning an image with startools and resize it with a classical image processor, given that I finally got the same scale/resolution on both cases.

Thanks!

Hi mguiterrez,
Apart from the specific technical solutions, one main difference is simply the spot in the workflow: Binning in ST is used on the virgin stack dataset PRIOR any post-production. This will increase the SNR and Bit Depth of the binned stack, increasing operational quality of all following modules. This is because any detail enhancement would not have to operate on so much noise and may focus on real detail instead, if binned. In contrast, resizing in a classical image processor is usually applied to the final image, hence it may only reduce any amplified noise and potentially created artifacts. Also, the image might have inferior quality because modules could not be pushed so hardly on the dataset.
clear skies

Thanks for the reply.
Technical and logically, is there any difference in the process itself?

mgutierrez wrote: ↑Thu Jul 22, 2021 3:29 pm Thanks for the reply.
Technical and logically, is there any difference in the process itself?

There is, indeed, a fairly big difference.
The purpose of binning is to "pool" co-located pixels into one, more precise "super" pixel (almost always by averaging the values, however other statistical outlier rejection methods can be applied as well).

Resizing does not have that goal, and is usually more concerned with preserving detail, sharpness or preventing aliasing.

Binning in StarTools averages co-located pixels when dealing with integers (2x2, 3x3) or alternatively (when dealing with fractional values) employs a "resizing" algorithm with an anti-aliasing filter that reduces overshoot (ringing) as much as possible. This filter was specifically designed for the purpose of keeping the input suitable for highly sensitive algorithms like deconvolution.

For more in-depth info, see the Bin module documentation.

Hope this helps!

great Ivo, thanks for the explanation!