search

hpkim0512 / Deep_MRI_Unet

17 stars 11 forks source link

No pre-trained model #2

Closed dedmari closed 6 years ago

dedmari commented 6 years ago

Is there any pre-trained model available?

hpkim0512 commented 6 years ago

Thank you for your attention. I uploaded the pre-trained model on README. Please check about that. Thank you

vegarant commented 6 years ago

Do you have the sampling pattern you used for the pretrained network?

hpkim0512 commented 6 years ago

We used a uniform subsampling of factor 4 (25% k-space data - 64 lines of a total 256 lines) with a few low frequencies (abot 4% k-space data - 12 lines of a total 256 lines). For further details, please see https://arxiv.org/pdf/1709.02576.pdf

vegarant commented 6 years ago

I was unsure if you started the sampling the at line 0,1,2 or 3, since I open the data in Matlab, and did a fft2 on it. The result was quite surprisingly non-sparse. However, using numpy.fft.fft2 gives a sparse result, so I can see that you started on line 0.

Thanks for the quit reply.