Hello-
To reiterate Jeff's answer, my dcm2nii is designed to go from DICOM to NIfTI and not vice versa. NIfTI is a simple, explicit lean image description whereas DICOM is a verbose, format that is so complicated it has been interpreted differently by different vendors. A proper DICOM requires a lot of information that simply no longer exists in the NIFTI format. A possible solution for this would be to convert NIfTI to a DICOM image with a modality of ‘OT’ (OTher) rather than ‘MR’ or ‘CT’. The OT type does not require sequence information that is lacking in NIfTI (e.g. field strength, TE, TR, etc). I think a larger problem is that most scanner manufacturers have implemented the DICOM format differently, and their viewers may struggle with reconverted data - e.g. should volume be stored as mosaics (Siemens), 2D slices (GE) or 3D volumes (Philips)? I think a very specific NIfTI to DICOM solution that supports a particular viewer would be very easy, but a general solution that supports the expectations of various viewers would be challenging. The 2D approach would be the most general but would need to compute slice angulation and offset correctly. I always tell people that DICOM is like the English language, that is used differently in England, US, NZ, Australia….
|