search

mylyu / pymrghost

python implementation of MRI EPI ghost correction, partial Fourier reconstruction, and other common functions.
MIT License
13 stars 4 forks source link

Not Working for standard Spin-Echo EPI #1

Open wenshangwang opened 3 months ago

wenshangwang commented 3 months ago

Dear Mengye

I am trying to handle the N/2 ghosting artifact in my Spin-Echo EPI image after reconstructed from k-space data. I put the data in the desired format (kx, ky, slice, coil) and run your code: image Where: image is my data with 128 readout samples, 64 PE lines, 26 slices and 44 channels. Which is one echo of the SE EPI.

What suggestion would you have for me to try. Thanks you very much

Wenshang

mylyu commented 4 weeks ago

Sorry for the late reply. Was your data parallel image accelerated? If so, the entropy based correction method won't work well. In such case, I usually acquire a fully sampled reference scan (e.g. multishot EPI, 3-echo navigator) to run the entropy based correction and apply the obtained parameters to the actual accelerated data. Alternatively, I recall there are some low rank based correction methods that can work with accelerated data.