search

carsonpurnell / cryotomosim_CTS

cryo-EM tomogram simulator
Other
19 stars 3 forks source link

Possibility to simulate VPP tomograms #16

Open frosinastojanovska opened 1 week ago

frosinastojanovska commented 1 week ago

Hello, is there a possibility of simulating the VPP effect with an added phase shift to the defocus?

I already created a large dataset of hundreds of tomograms to train the DL models I was developing. Yet, in the real dataset, I find tomograms imaged with VPP, and I need simulations that can also help identify the accuracy of the different versions of tomograms (only defocus or defocus + phase shift). Let me know what you think about this :).

carsonpurnell commented 1 week ago

I am interested in getting phase contrast working, but I do not myself use it. Would you be able to send me a handful of example tomograms of each, and your imaging conditions? Most of the necessary backend already exists, I just need to properly calibrate and provide access to its parameters. Proper data will help with that.

frosinastojanovska commented 6 days ago

Hi @carsonpurnell, this can be a nice dataset for VPP, since it has the same sample imaged with only defocus and with VPP: https://www.ebi.ac.uk/empiar/EMPIAR-10988/

carsonpurnell commented 6 days ago

I've published a new release version with very basic phase contrast support. the simulation parameter is 'phase' and input value in radians. the contrast spectra looks reasonable, but the SNR and resolution is not accurate out of the box. You will need to manually adjust other parameters, especially dose, to get a more realistic simulation if you run at pi/2 phase. Let me know how it works out, and how it compares to your datasets for later improvements.