How To - TiffStackReader, custom Geometry, Normaliser

[M-A-Demir]

Added TiffStackReader.ipynb how-to to the '1_Read_and_visualise' folder.
Added 2_Geometry/CreateCustomGeometry.ipynb
Added 3_Processors/FlatDarkFieldNormaliser.ipynb

[M-A-Demir]

Potential issue: TIFFStackReader.ipynb downloads the sandstone dataset to its directory (how-tos/1_Read_and_visualise), CreateCustomGeometry.ipynb uses the already-downloaded dataset, but file path was not updated in the notebook.

Where should this dataset be downloaded to in this repo, so I can adjust the file path variables accordingly?

[hrobarts]

Hi Mariam, these are looking really great - I think they're all very clear. I couldn't add comments on the code for the TIFFStackReader and FlatDarkFieldNormaliser so putting some here:

• In the TIFFStackReader you could say it accepts .tiff and ,tif files
• In the FlatDarkFieldNormaliser you could use show2D([sandstone, sandstone_norm]) to show the plots next to each other, you could add titles as well. Maybe the same for the flat and dark plots
• I think the %store looks like a good way to get the previously processed data but we should discuss how that would work if we're going to use GHA to render these

[hrobarts]

I think this simple geometry example will be really useful for people. I wonder if we should have a separate 2D cone beam, 2D parallel beam, 3D cone beam and 3D parallel beam example?

[MargaretDuff]

Hi @M-A-Demir, thanks for your hard work on this! I think all the code is there but there are some places where we could improve the explanations.

• In the FlatDarkFieldNormaliser notebook, it would be great if you could explain at the top of the notebook what a flat and dark field is and why you would correct for it.
• In the CreateCustomGeometry notebook, where did you get the SrcToObject and SrcToDetector values from? It might be worth giving a little more explanation of where you found these but mentioning that it will be different for different systems and ways of saving data.
• In the TiffStackReader notebook, you create a list of tiff_files excluding some of the BBii_????.tif files. Perhaps point to the normalisation notebook with your explanation as to why?

I will leave a few more specific comments in the notebooks

[MargaretDuff]

Not sure this sentence makes sense

[MargaretDuff]

Might also be worth showing an FBP reconstruction, to demonstrate that the geometry is correct and thus gives a reasonable reconstruction.

[M-A-Demir]

Thanks Margaret! I added the clarifications you suggested, as well as the reconstruction (& CoR adjustment) to the geometry notebook

[MargaretDuff]

Sorry Mariam, I made a few suggested changes and then between vscode and github they all got deleted. Will try again another time

[MargaretDuff]

Thanks @M-A-Demir. The code is all there. I have made some suggestions to improve the readability!

[MargaretDuff]

What does this do? Generally users will take notebooks verbatim and try and adapt them to their dataset so if this is a cell particular to the how-tos and a not general CIL recommendation then we should state that.

[MargaretDuff]

Is this needed in this PR?

[M-A-Demir]

That file is only there so I could push the 'data' directory, as I couldn't figure out how to push an empty directory

[MargaretDuff]

Does this mean you have to run the previous how-to first. This needs to be made clear if so!

[hrobarts]

Hi Mariam, these are all looking great. I'd be happy to go ahead and merge the TiffStackReader and Normaliser notebooks. I think we should add a simulated example to the Normaliser in future but this works well as a real example.
I would maybe ask @gfardell her opinion on the geometry notebook