search

OpenChemistry / tomviz

Cross platform, open source application for the processing, visualization, and analysis of 3D tomography data
https://tomviz.org/
BSD 3-Clause "New" or "Revised" License
325 stars 87 forks source link

Question: camera rotating around object #1871

Open klaheeman opened 5 years ago

klaheeman commented 5 years ago

Can I reconstruct an object when the detector is rotating around the object instead of the object tilting/rotating around an axis?

Clearly some geometry parameters would be needed: source-detector distance, projection of the source position on the detector, and the x-y-z position of the fixed object with respect to the rotation axis.

Best would be to have a UI-based setup, but also fiddling with a script would be ok for me. Is this already possible? It would be ok for me to do some changes to the source code as well to try out this new geometry.

Being TomViz such a nice a complete tool for tomography, I think it would be good to include also the "medical-style" rotating detector geometries.

cryos commented 5 years ago

You may have to fill me in on what you think is missing. If the detector is rotating around the sample, but the rest of the geometry remains the same then can't you just attach a tilt angle equivalent to the detector rotation? If the geometry is more exotic then it would entail adding more/new metadata to a tilt series that contained this information.

We store tilt angles in tilt series, but there could be other image projection series where each projection had a larger geometry object. The reconstruction algorithms would also need to be rewritten to account for them.

ercius commented 5 years ago

Another issue with medical style tomography is that they use fan beam rather than parallel beam projection geometries. I'm not sure if this is also part of the original question.

On Mon, Jun 24, 2019, 11:12 AM Marcus D. Hanwell notifications@github.com wrote:

You may have to fill me in on what you think is missing. If the detector is rotating around the sample, but the rest of the geometry remains the same then can't you just attach a tilt angle equivalent to the detector rotation? If the geometry is more exotic then it would entail adding more/new metadata to a tilt series that contained this information.

We store tilt angles in tilt series, but there could be other image projection series where each projection had a larger geometry object. The reconstruction algorithms would also need to be rewritten to account for them.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/OpenChemistry/tomviz/issues/1871?email_source=notifications&email_token=ADXORFAYKGGDM7WSCTPR5ULP4EFBDA5CNFSM4H3ANJWKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYNYLXQ#issuecomment-505120222, or mute the thread https://github.com/notifications/unsubscribe-auth/ADXORFEMDOU5TSRY3A7WTO3P4EFBDANCNFSM4H3ANJWA .

klaheeman commented 5 years ago

All correct. I need a cone beam geometry and indeed the rotation of the object vs. the rotation of the detector assembly does not make a difference.

So all in all I would like support for cone beam geometry: source detector distance, position of the Xray focal point.

Cone beam tomography support would open up tomviz to industrial and medical tomography. The medical setup which I'm interested in would typically require a rotation around the object. Calibration is always needed. For medical also helical wide fan beam would be good, but those are expensive devices and people typically already have all the tools required. For industrial, sparse views are sometimes used to capture large objects.

Is anybody interested in adding cone beam support to tomViz(FDK or TV)?

thanks,

C.

Il giorno mar 25 giu 2019 alle 00:15 Peter Ercius notifications@github.com ha scritto:

Another issue with medical style tomography is that they use fan beam rather than parallel beam projection geometries. I'm not sure if this is also part of the original question.

On Mon, Jun 24, 2019, 11:12 AM Marcus D. Hanwell <notifications@github.com

wrote:

You may have to fill me in on what you think is missing. If the detector is rotating around the sample, but the rest of the geometry remains the same then can't you just attach a tilt angle equivalent to the detector rotation? If the geometry is more exotic then it would entail adding more/new metadata to a tilt series that contained this information.

We store tilt angles in tilt series, but there could be other image projection series where each projection had a larger geometry object. The reconstruction algorithms would also need to be rewritten to account for them.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub < https://github.com/OpenChemistry/tomviz/issues/1871?email_source=notifications&email_token=ADXORFAYKGGDM7WSCTPR5ULP4EFBDA5CNFSM4H3ANJWKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYNYLXQ#issuecomment-505120222 , or mute the thread < https://github.com/notifications/unsubscribe-auth/ADXORFEMDOU5TSRY3A7WTO3P4EFBDANCNFSM4H3ANJWA

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/OpenChemistry/tomviz/issues/1871?email_source=notifications&email_token=AHU6QM7JXCBQV7JCAN6ROMDP4FBQRA5CNFSM4H3ANJWKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODYOML4A#issuecomment-505202160, or mute the thread https://github.com/notifications/unsubscribe-auth/AHU6QMZNAZGA4ZYFDP5WUVTP4FBQRANCNFSM4H3ANJWA .