Ozan Öktem

I think [Euler angles](https://en.wikipedia.org/wiki/Euler_angles) Is just another way to parametrise a 3D parallell beam geometry commonly used in both electron tomography and single particle analysis.

The note [FBP_in_3D.pdf](https://github.com/odlgroup/odl/files/3234520/FBP_in_3D.pdf) derives explicit expressions for the filter in FBP for inverting the ray transform restricted to a general parallel beam line complex in R^n. One may also consider...

This has been discussed at several occasions and especially so for algorithms that want to make use of automatic differentiation in PyTorch. A closely related issue is #731 (see also...

I also think the detector collimators are slanted in tomosynthesis and the source moves along a linear trajectory. At least hat was the case with the X-Ray detector for tomosynthesis...

Somewhat off-topic, but not entirely: @wjp , what happened to the initiative for introducing native support for curved detectors (cylindrical and spherical) in ASTRA? Just curious. PS: I am aware...

@JevgenijaAksjonova or @sbanert, do you have any input regarding the mathematical question that @mehrhardt posted?

To clarify, freely available software packages for computing ray transform and corresponding backprojection (like [ASTRA](https://www.astra-toolbox.com), [RTK](https://www.openrtk.org), [CASToR)](http://www.castor-project.org) do not support 3D helical CT scanning with curved detectors. ODL has data...

Adopting the ODL operator centric viewpoint, the interpolation from curved to flat detector is (similar to rebinning) a mapping between two geometry objects. So why not have it as an...

@JevgenijaAksjonova , here is a suggestion on how to proceed: 1. Make sure ODL code for ray transform includes a check whether the back-end supports curved detectors. 2. Define interpolation...

There seems to be a scaling issue in how FFT is used in the context of FBP.