Revision as of 19:12, 3 February 2009

Home < ChangeTracker

Overview

THIS PAGE IS UNDER CONSTRUCTION -- CHECK BACK OFTEN!

ChangeTracker is a 3D Slicer GUI module for monitoring slowly evolving tumors. This page will host the relevant information about ChangeTracker.

Jan 25, 2009, use of 3d box widget with tumor volume rendering for ROI selection
Jan 12, 2009, New visualization features integrated
Jan 4-9, 2009, ChangeTracker project on NA-MIC AHM/Programming Week at Salt Lake City. See summary and results here
Dec 27, 2008, DiameterEstimator tool is available for automatic estimation of tumor diameter from the binary image

during "Analysis type step", Slicer crashes if "Analyze" button is hit more than once
registration of a new volume node with the scene results in update of the slice views: Resolved: this appeared to be due to a lacking functionality in SliceCompositeNode, which invoked PropagateVolumeUpdate() on image load. This functionality can now be controlled via the DoPropagatVolumeUpdate in SliceComposite MRML node
incorrect setting of the registration parameters results in registration producing empty volume, no error messages from the module: Resolved: committed a fix to CommandLineModuleLogic to update the module node status correctly to indicate the error, if module entry point returns non-zero. TODO: update ChangeTracker logic to account for this change
need to investigate reproducibility of registration and impact of the random seed initialization: Linear registration updated with metric seed reinitialization
investigate reasons for segfault with the weird dataset from Yanling: Resolved (?): data is intrinsically very difficult to align. Registration must be solved prior to using ChangeTrcker
check and clean up file output procedures (..-TG directories)

multi-color volume rendering for tumor growth results: DONE
use Compare view: DONE
3d box widget for ROI selection DONE
analysis of multiple time points
testing IN PROGRESS
improved tumor segmentation to exclude prominent non-tumor regions (vessels, scull, falx)
Jean-Jacques Lemaire: would be nice to combine growth results with the 3D rendering of the estimated tumor volume
Jean-Jacques Lemaire: improve user interface for tumor boundary identification -- something missing in the existing tools; would like to see better robustness of segmentation
Ron wish list: get user input for inside/outside locations of points for tracked structure, use this for threshold selection and initial segmentation; ROI centering; sync ROI scales colors with the 3d box colors; review naming conventions for intermediate volumes; allow for separate show/hide functionality for 3d box and volume rendering; link compare views; support scene snapshots; fix bottom layout panel random popup issue
graceful error handling

E.Konukoglu, W.M.Wells, S.Novellas, N.Ayache, R.Kikinis, P.M.Black, K.M.Pohl. Monitoring Slowly Evolving Tumors. Proc. of 5th IEEE International Symposium on Biomedical Imaging: From Nano to Macro 2008, pp.812-815 link

Please contact Andriy Fedorov for further information.

@@ Line 30: / Line 30: @@
 * investigate reasons for segfault with the weird dataset from Yanling: ''Resolved (?): data is intrinsically very difficult to align. Registration must be solved prior to using ChangeTrcker''
 * check and clean up file output procedures (..-TG directories)
-* If the image dimension extends in negative subspace, for example, L-R extent is [0,-400], the scale range during ROI selection will not be displayed properly. This is a known KWWidgets [http://www.kwwidgets.org/Bug/view.php?id=8458 bug]
 =Feature requests/Features under development=