Hi,
I’d like to output images to be read into different software for analysis. Is there a way to do this using a standard format like nifti or dicom?
Shaihan
Hi Shaihan!
I’ve made a notebook to convert kspace data files in ‘.npy’ format to image data files in nifti format. If you update to the latest version of the ilumr software (2.2.0), the nibabel library will be installed and you can run the notebook on the system. Just drag it into the jupyterlab file browser to upload it, then double click to open.
nifti_convert.ipynb (3.9 KB)
For more info on the software update see: https://www.resonint.com/post/ilumr-software-update
The notebook will convert a whole directory of files at once, from the IMPORT_DIR, and save them to the EXPORT_DIR. There are also some image processing options.
You can also run the notebook on your PC if you have python and jupyterlab installed. You may need to install the numpy and nibabel libraries yourself:
1 Like
It would be cool if:
(a) the k-space data could be saved in a DICOM Raw Data SOP Class instance
(b) reconstructed images could be saved as DICOM MR Image or Enhanced MR Image (multi-frame) SOP Class instances (or both)
in both cases with the metadata populated with the acquisition parameters that describe the pulse sequence (which is the point of using DICOM) in addition to the spatial information.
If the goal is to teach MRI, it would be cool to teach not just the physics, but the real-world use (which in clinical medical imaging means DICOM, not research formats like NIfTI).