MRI Partial Fourier zero filling reconstruction - image aspect ratio problem.

5 次查看（过去 30 天）

Roshtha 2018-1-12

1
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/376800-mri-partial-fourier-zero-filling-reconstruction-image-aspect-ratio-problem

编辑： Lauren Hirt 2018-1-25

Hello all,

I have a partial fourier k-space data of size 512x189 (ROxPE) from siemens machine. The reconstructed MRI DICOM image from the scanner has 512 rows and 432 columns.

I tried zero filling method for reconstructing images from k-space.

Created a matrix of size 512x432.
Assigned first 189 PE lines with acquired data and rest set to zero
Fourier transform to reconstruct images
As shown in the images below, the aspect ratio of reconstructed image is lost with zero filling.

I also tried with PE dimension size set to next power of 2 ie. 512. But still the aspect ratio of image is not preserved.

How can I reconstruct images with increased size?

Thanks.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Lauren Hirt 2018-1-25

编辑：Lauren Hirt 2018-1-25

Could you post the code you're using to ZIP (zero-fill) your data? Could be an issue there.

Also, I'm not sure where your data came from (professor, download, research machine), but remember that most of the time DICOM images have been significantly processed both before and after kspace. I've compared slices from kspace data that showed slightly different images from the associated DICOM files, as if they were a different slice thickness. DICOM aren't terribly helpful to image processing researchers since they no longer contain real and imaginary components and have been heavily pre-processed. Often you don't know what has been preformed on them either. Just something to consider :)

请先登录，再进行评论。

请先登录，再回答此问题。