search

cni / nims

Legacy version of NIMS. All new development at @scitran
14 stars 12 forks source link

dicom slice order #48

Closed rfdougherty closed 9 years ago

rfdougherty commented 11 years ago

The dicom slice order calculation here: https://github.com/cni/nims/blob/nimsdata/nimsdata/nimsdicom.py#L255 can fail if a scan was aborted and the last dicom happens to have the same position as the first. It would be better to just look at the slices for one volume. In fact, we should refactor so that all the metadata computations can work on just one volume of dicom slices. Then, if you just need metadata, you don't have to load a whole time series.

gsfr commented 11 years ago

Is this still an ongoing issue?

rfdougherty commented 11 years ago

yes, but it's an edge case, so low priority right now.

kevlarkevin commented 9 years ago

I think @rfdougherty fixed for GE already. The GE slice order calculation uses trigger_times, instead of position. The GE slice order calculation only looks at one volume.

The Siemens slice order is stored as a code in the header, and then translated to the NIFTI slice order code.