How to add a PSF to a Multiview reconstruction dataset

fiji
imagej
deconvolution
multiview
psf

#1

Dear all,

I am trying to use Multiview-Reconstruction to register, deconvolve, and fuse a 5 angles SPIM dataset.

I don’t have beads with my samples. I do have, however, an independent tiff stack with beads that I could use for PSF extraction. The issue is that I am so far unable to add that tiff file to my Multiview xml dataset file.

At some point I thought that I should start the xml file already containing the SPIM stack plus the stack with beads for PSF extraction. But I cannot convince the plugin to load them together. The SPIM stack has several angles and timepoints, the psf stack does not. I am assuming that this difference in dimensions is getting in the way of loading them together on the same xml file.

Now, for the questions:
How do I add files (tiff stacks) to an already existing xml dataset?
Is it possible to have a psf from another xml dataset loaded to an existing xml dataset?

Does anyone know how to handle this one?

Thank you in advance,

Rafael


#2

Hi!

I manage to find a way to solve half of the problem.

One can create another xml project just for the PSF. In this one, the image stacks to generate the PSF are loaded, the PSF is extracted, and then one can save a tiff stack file of the PSF itself. This is done on the same menu as “extract PSF”, but one chooses to display it. That generates a new window in Fiji containing the PSF. This can be saved as any other tiff stack.

Back to the xml project of the SPIM dataset of interest, in the same menu to assign a PSF, one chooses “Advanced” and chooses to add a new PSF. That prompts the user to choose a file. I chose the tiff stack containing the PSF generated above.

I am not sure if this is the best way or even if it is strictly correct. I am still testing results. All I know so far is that the plugin accepts that tiff stack as a PSF and performs a deconvolution plus fusion of the 5 angles. I still get some artifacts on the final fused image, but I cannot tell if they come from the data, from sub-optimal PSF generation, from wrong use of the plugin, or else.