mirp has a flexible system for determining which files are images and masks are available. It matches those files and masks based on various heuristics.

The efficiency of this process can be improved, particularly for DICOM files.

To identify and sort through DICOM data, information is read from DICOM headers using pydicom. This is a relatively time-consuming affair, even though only few tags are relevant.

This would also be relevant for other file types. However, we don't have that much influence on reading information from other file types, since these rely on itk.