Question about analytical reconstruction for helical cone-beam CT

[zezisme]

Dear CIL developers,

thank you very much for developing and maintaining the CIL reconstruction framework. It is a very impressive and useful platform.

I would like to ask a specific question regarding analytical (non-iterative) reconstruction support for helical cone-beam CT:

• Does CIL currently support analytical or approximate analytical reconstruction for helical (spiral) flat-panel cone-beam CT,
  such as helical FDK-type methods, rebinning-based FDK, or other approximate analytic approaches?
• Or is the current cone-beam support limited to circular trajectories, with helical geometries handled only via iterative reconstruction?

My interest is specifically in FBP/FDK-like analytical reconstruction, not iterative algorithms (e.g., SART, TV-based, etc.).

If helical analytical reconstruction is not currently supported, I would also appreciate knowing:

• whether there are plans to support it in the future, or
• whether the recommended workflow would be to perform helical-to-circular rebinning externally and then use CIL’s FDK implementation.

Thank you very much for your time and for making CIL available to the community.

Best regards

[gfardell]

Hi @zezisme,

Thank you for you lovely comments, it's always nice to here from CIL users!

Currently, the Cone3D_Flex geometry allows the standard projectors to handle non-standard trajectories, but as you spotted, this mostly benefits iterative methods . cil.plugins.Astra.FBP will accept the geometry, but it will apply a standard FDK reconstruction without accounting for the helical pitch.

We don't have any analytical helical algorithms implemented at the moment, and they are not on the immediate roadmap. However, we know there is interest in this area, and we’d love to learn more about your specific use case.

Contributions are very welcome, and we can work with you if you’d like to explore approximate methods via rebinning or pre-filtering and weighting. Implementing modified FDK methods is more involved since they require projectors that incorporate helical cone-beam weightings, but it’s certainly possible if there’s demand.

For now, the recommended workflow would be to perform helical-to-circular rebinning externally and then use CIL’s FDK implementation. Please join our discord community and let us know how you get on, we're very keen to have a helical examples in our CIL-User-Showcase.

Thanks again for reaching out, and please let us know more about your application, we’re always keen to support new developments!