Releases: rordenlab/dcm2niix
Releases · rordenlab/dcm2niix
version 7-February-2017 (v1.0.20170207)
- Can be compiled to use either Philips Float or Display intensity intercept and slope values.
- Handle 3D Philips DICOM and PAR/REC files where images are not stored in a spatially contiguous order.
- Handle DICOM violations where icon is uncompressed but image data is compressed.
- Best guess matrix for 2D slices (similar to dcm2nii, SPM and MRIconvert).
- Linux (case sensitive filenames) now handles par/rec as well as PAR/REC.
- Images with unknown phase encoding do not generate BIDS entry.
- Unified printMessage/printWarning/printError aids embedding in other projects, such as divest.
For compiled code, download the MRIcroGL package for MacOS, Linux or Windows.
version 30-January-2017 (v1.0.20170130)
- Handle 3D Philips DICOM files where images are not stored in a spatially contiguous order.
- Handle DICOM violations where icon is uncompressed but image data is compressed.
- Best guess matrix for 2D slices (similar to dcm2nii, SPM and MRIconvert).
- Linux (case sensitive filenames) now handles par/rec as well as PAR/REC.
- Images with unknown phase encoding do not generate BIDS entry.
- Unified printMessage/printWarning/printError aids embedding in other projects, such as divest.
Version 1-Nov-2016
- AppVeyor Support (Ningfei Li & Chris Filo Gorgolewski)
- Swap 3rd/4th dimensions for GE sequential multi-phase acquisitions (Niels Janssen)
version 30-Sept-2016
30-Sept-2016
- Save ImageType (0x0008,0x0008) to BIDS.
- Separate CT scans with different exposures.
- Fixed issues where some compilers would generate erratic filenames for zero-padded series (e.g. "-f %3s")
- Reduce verbosity (reduce number of repeated warnings, less scary warnings for derived rather than raw images).
- Re-enable custom output directory "-o" option broken by 30-Apr-2016 version.
- Deal with mis-behaved GE CT images where slice direction across images is not consistent.
- Add new BIDS fields (field strength, manufacturer, etc).
- Philips PAR/REC conversion now reports inconsistent requested vs measured TR (due to prospect. motion corr.?)
- GE: Locations In Acquisition (0054, 0081) is inaccurate if slices are interpolated, use Images In Acquisition (0020,1002) if available
- New filename options %d Series description (0008,103E), %z Sequence Name (0018,0024)
- New filename options %a antenna (coil) number, %e echo number
- Initialize unused portions of NIfTI header to zero so multiple runs always produce identical results
- Supports 3D lossless JPEG saved as multiple fragments