User:Millerjv/ProvenanceWorkflow
Jim, Harini, and Steve had a tcon on 2012-03-08 to discuss extensions to the Editor. The discussion splintered to a variety of topics that all pointed back to the need to associate datasets together. The motivations are twofold. First, we are expecting a lot of user to mentally track the relationships between volumes, e.g. these two volumes are from the same subject, these two are from the same series, this volume is derived from that volume. Second, there are times where entire groups of data need to be managed in concert, e.g. multiparametric sequences passed into modules, etc.
Currently, this is done is two ways in Slicer. First, by node naming conventions, we track relationships between data. Second, by using HierarchyNodes to define explicit relationships.
Class of problems that require an association between datasets:
- Organizing nodes based on Patient-Study-Series
- Could be utilized when switching to CompareView and automatically creating the proper number of viewers and populating the background.
- Associating derived data with the source data
- Associate CLI outputs as derived data from CLI inputs
- Record "macros" and workflows
- Grouping multivariate data (multiple volume nodes) with the Editor
- Grouping the "training" and "result" data used in GrowCut with the original volume
- Extending the NodeComboBox to group nodes into "Recent", "Associated", and "All" subgroups
- Useful when the Editor is open and the user selects different label maps in the Red Viewer
Extensions also discussed
- Allow the LabelMap ComboBox in the SliceControllerWidget to create a new label map, defaulting to the properties of the background volume in the same Slice Viewer.
