Skip to content

SPMIC-UoN/fproc

Repository files navigation

File pre-sorter for imaging processing pipelines

A common feature of most imaging processing pipelines is an initial step where input data (in DICOM or NIFTI) format is identified for particular uses in the pipelines. For example, we may be expecting to find a T1 map, a T2* weighted scan and a set of DWI volumes. Usually this identification involves an ad-hoc mixture of filename matching and metadata inspection.

Fsort is intended to be a configurable drop-in process capable of performing these kind of matching tasks and writing output files to a folder with standardized naming. However the semantics of the matching - how we identify the T1 map - are part of the configuration and need to be tailored for each application.

An important principle of Fsort is that filenames should never used in matching. DCM2NIIX creates filenames based on metadata and the metadata should be inspected directly as that is more reliable than parsing filenames with globs and regular expressions. However some basic filename matching functionality is available for use cases where, for example, only NIFTI data is availabel and no JSON metadata sidecar is provided.

About

Base for simple processing pipelines

Resources

License

Stars

Watchers

Forks

Packages

No packages published