search

ketanfatania / QMRI-PnP-Recon-POC

Plug-and-Play Magnetic Resonance Fingerprinting based Quantitative MRI Reconstruction using Deep Denoisers (Proof of Concept) (IEEE ISBI 2022)
14 stars 4 forks source link

Using your algorithm for 3d reconstruction #1

Closed jakubMitura14 closed 5 months ago

jakubMitura14 commented 2 years ago

Hello thank you for publishing and sharing your work I am wondering weather i could use your algorithm for 3d reconstruction of classical Mri

I mean for example i have three series transverse t1, saggital t1 and coronal t1 and using all of those i would like to reconstruct the single series with maximized spatial resolution, this obviously requires sampling and dealing with noise - can your algorithm can be used for it ?

Thank you !

ketanfatania commented 2 years ago

Hi Jakub,

Thank you for your interest in our research

The PnP-ADMM algorithm can absolutely be used for conventional MRI. Here is a link to one of several previous works which has applied PnP-ADMM to conventional MRI https://doi.org/10.1109/MSP.2019.2949470 .

I do not know much about 3D reconstruction for MRI but will try to answer your question as best as I can. From my understanding, it seems that 3D reconstruction of MRI uses the axial, sagittal and coronal slices of a scan to create a 3D MRI via some image processing method (please correct me if I am wrong). Based on this, I think PnP-ADMM could be used to improve the quality of each 2D slice within each plane. This may then improve the quality of the 3D MRI after it is created. In this case you may need to train three separate denoisers, one for each plane.

Please let me know if you have any other questions or would like to discuss further. If you decide to use our code, I would be happy to provide support if there are any issues or questions.

I hope that helps

jakubMitura14 commented 2 years ago

All agrees, thanks for detailed answer ans support offer !

I have now idea how to use it as you described, I will work on it and probably come back thanks !

ketanfatania commented 2 years ago

Apologies for the late reply

No problem!

All the best for your work

jakubMitura14 commented 2 years ago

If you will allow, I will ask further - so does this algorithm can be used to dicom images (I want to denoise and if possible resample them using your algorithm to improve later interpolation between different modalities in prostate MRI) , or it requires reengineering :)? secondly as I do not have matlab I will need to use octave did you tested it in octave

ketanfatania commented 2 years ago

It should be possible to use the PnP-ADMM algorithm for DICOM images. You'll have to modify our code to accept DICOM images. In regards to Octave, I have not tested our code with it and am not sure if you'll have issues running our code with it.

Hope that helps

jakubMitura14 commented 2 years ago

Thanks!!