Release Note
- TopK issue patch
Do you need any help?
If you have any issues or have questions about this model or the Siril's script, please open an issue in this repository.
For common issues that may arise I am going to keep updated this page for fixes.
- Computers with Intel Ultra may face issues running this model.
How to use
This model has to be used in conjunction with a script. I officially support AberrationRemover script in Siril, but SetiAstro provides one for SASpro. In Siril the script has to be downloaded by going in the top bar > scripts > get scripts > AberrationRemover.
This model is intended for linear images after background removal and color correction.
For best results, apply it before any deconvolution or noise reduction steps.
Starting from version 1.1.0 of Siril's script, new parameters are available:
- strength: controls the intensity of the aberration correction.
- Dark ringing reduction: applies an additional filter to suppress dark ringing artefacts that may appear around bright stars as a side effect of strong correction. This option should be enabled only if such artefacts are visible.
- Protect Background: limits the application of AR to non-background pixels, preventing over-processing in low-signal areas.
What does this do?
Aberration Remover (AR) is an AI-based tool designed to remove optical aberrations from images, affecting both stellar and non-stellar components. It can correct monochromatic aberrations such as defocus, astigmatism, coma, trefoil, spherical aberration, and other higher-order aberrations.
AR is not designed to correct sensor or optical tilt, lateral chromatic aberration, or to compensate for atmospheric seeing.
AI model licensing
This AI model is released under the Creative Commons Attribution–NonCommercial–ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.
The following points are not legally binding and are provided for clarity only.
For the authoritative terms, consult the LICENSE.md file.
You may:
- Use, modify, and fine-tune the model for non-commercial purposes.
- Create derivative models, including fine-tuned versions or adaptations.
- Train other models using this model, in whole or in part, provided the result is a derivative work.
- Redistribute the original or modified model, as long as the license terms are followed.
You must:
- Provide proper attribution to the original author(s).
- License any derivative work (including fine-tuned or trained models) under CC BY-NC-SA 4.0.
- Include a copy of the license and indicate if changes were made.
You may not:
- Use the model or derivatives for commercial purposes, including but not limited to:
- Paid software or services
- Subscription-based products
- Software with ads or other monetization
- Use within commercial or revenue-generating workflows
- Re-license the model or derivatives under more permissive or proprietary terms.
- Circumvent the non-commercial restriction by indirect or bundled commercial use.