Back to Projects List
Slicer Case Iterator
Key Investigators
- Joost van Griethuysen (The Netherlands Cancer Institute)
- Steve Pieper (Isomics)
Project Description
This project aims to expand the usage of Slicer Case Iterator to work with DICOM(WEB) backend to support workflows where the data is available in DICOM format.
Objective
- Generalize internal batch definition to allow different input modes (CSV, DICOM)
- Link to Slicers DICOM browser to retrieve images
- Implement prefetching of next n cases to speed up processing
- Store segmentation results as DICOM SEG objects
Approach and Plan
- Create class implementing generator to hold batch, with different input methods a inherited subclasses
- Talk to Slicer developers on how to best approach this (what kind of identifiers, methods to use, etc.)
- Closely related to point above. Investigate on how to analyze input to determine best conversion step, how to store results of this evaluation.
- Talk to Slicer developers on how to store a Slicer Segmentation object as DICOM SEG object (when input image is DICOM format)
Progress and Next Steps
- Done, merged with master in PR #7
- Using SeriesUID, it is now possible to load series using the Slicers DICOM database. Not yet merged with the master PR #8
- In DICOM input, series is first checked for loadables using MultiVolumeImporterPlugin and if necessary, by DICOMScalarVolumePlugin. Not yet possible to do this using a type of prefetch, or parallel processing.
Next Steps
- Add possibility to store Segmentations as DICOM SEG objects
- Add example scripts to help users generate the input tables for their datasets
- Investigate use of DICOM Web UPS / Modality Work List to define the batch
Illustrations
Background and References