Would love some help finetuning high rejection during Image Integration

Remco Hekker

After integrating my lights, I noticed that some large stars had nasty cerated edges around them.
These are not present in the original light frames and they leave bad artifacts the more I process my image.
(The color fringing itself is handled very well by the new version of BlurXterminator BTW.)

After inspecting the rejection_high map, I think I found the problem. 
But since the halos are brighter in the rejection map than some of the other stuff that needs to be rejected, I'm not sure where to look for improvement.
It's a stack of 575 lightframes so random experimentation takes a lot of time! I'm currently using GESD with the default settings of: Outliers 0.30 ESD significance of 0.05

Any help will be greatly appreciated.

Kind regards, 
Remco

Remco Hekker

I realized the previous image might not show the problem very well. So I'll share a better example.

Remco Hekker

For anyone who has is interested in this topic:
It turns out the video: "Optimizing Rejection: Minimizing Artifacts" Adam published earlier this year covers exactly this situation. The new version looks a lot better!