oymd avatar

Hi everyone

I am struggling to avoid removing the small fuzzy nebulae from my galaxies alongside the stars when using the recent AI star removal tools.

With OSC broadband data I usually go through gradient removal, then blurX Correct only, then SPCC, then blurX, then mild NoiseX, then subtract stars before using MST stretching, then work on the non linear starless galaxy image, and in the end stretch the stars separately and combine into a final image.

My latest projects have been M63, and more relevant to the matter, M101.

My issue is that starX completely removes all the small red nebulae scattered all across the galaxies arms, in addition to the stars.

I tried using masks but that did not help.

tried LARGE OVERLAP and same results.

The problem is that both StarX and Starnet++ do not have any settings or thresholds one can tweak or play with to avoid this issue.

I always prefer to remove the stars before stretching and I am stuck.

Is there a trick I am missing?

I even tried to avoid blurX early in the processing to avoid sharpening the nebulae, so as starX would leave it alone. It STILL removes the blurry nebulae alongside the stars??

I am surprised that the AI tools are so clever in identifying non stellar structures, as is the case with BlurX etc when enhancing them, yet so dumb in removing very fuzzy blurry nebulae alongside perfectly round stars in the case of StarX or Starnet++?

Any advice would be great.

Well written Respectful Engaging
AstroGadac avatar

oymd · May 21, 2026, 07:34 PM

Hi everyone

I am struggling to avoid removing the small fuzzy nebulae from my galaxies alongside the stars when using the recent AI star removal tools.

With OSC broadband data I usually go through gradient removal, then blurX Correct only, then SPCC, then blurX, then mild NoiseX, then subtract stars before using MST stretching, then work on the non linear starless galaxy image, and in the end stretch the stars separately and combine into a final image.

My latest projects have been M63, and more relevant to the matter, M101.

My issue is that starX completely removes all the small red nebulae scattered all across the galaxies arms, in addition to the stars.

I tried using masks but that did not help.

tried LARGE OVERLAP and same results.

The problem is that both StarX and Starnet++ do not have any settings or thresholds one can tweak or play with to avoid this issue.

I always prefer to remove the stars before stretching and I am stuck.

Is there a trick I am missing?

I even tried to avoid blurX early in the processing to avoid sharpening the nebulae, so as starX would leave it alone. It STILL removes the blurry nebulae alongside the stars??

I am surprised that the AI tools are so clever in identifying non stellar structures, as is the case with BlurX etc when enhancing them, yet so dumb in removing very fuzzy blurry nebulae alongside perfectly round stars in the case of StarX or Starnet++?

Any advice would be great.

In pixinsight you can mask the galaxy and apply starX to only everything else. It will however leave out the stars in the foreground of the galaxy too.

oymd avatar

AstroGadac · May 21, 2026, 08:15 PM

oymd · May 21, 2026, 07:34 PM

Hi everyone

I am struggling to avoid removing the small fuzzy nebulae from my galaxies alongside the stars when using the recent AI star removal tools.

With OSC broadband data I usually go through gradient removal, then blurX Correct only, then SPCC, then blurX, then mild NoiseX, then subtract stars before using MST stretching, then work on the non linear starless galaxy image, and in the end stretch the stars separately and combine into a final image.

My latest projects have been M63, and more relevant to the matter, M101.

My issue is that starX completely removes all the small red nebulae scattered all across the galaxies arms, in addition to the stars.

I tried using masks but that did not help.

tried LARGE OVERLAP and same results.

The problem is that both StarX and Starnet++ do not have any settings or thresholds one can tweak or play with to avoid this issue.

I always prefer to remove the stars before stretching and I am stuck.

Is there a trick I am missing?

I even tried to avoid blurX early in the processing to avoid sharpening the nebulae, so as starX would leave it alone. It STILL removes the blurry nebulae alongside the stars??

I am surprised that the AI tools are so clever in identifying non stellar structures, as is the case with BlurX etc when enhancing them, yet so dumb in removing very fuzzy blurry nebulae alongside perfectly round stars in the case of StarX or Starnet++?

Any advice would be great.

In pixinsight you can mask the galaxy and apply starX to only everything else. It will however leave out the stars in the foreground of the galaxy too.

Yes, tried that, which leaves all the stars within M101, and then their clors go all wonky when I am doing curves etc to the galaxy…

andrea tasselli avatar

Don’t remove the stars, nothing wrong with that. Not that it really matters if you stretch properly…