CIBC:Project:ImageVis3D:Meetings:2008-02-14
From NCRR Biomedical Software Development, Engineering, and Dissemination Wiki
ImageVis3D Meeting
February 14
Allen, James, Jeroen, Dave, Rob, David
- Goals for next time:
- key differntiating features
- multidimensional transfer functions for great volume rendering
- multifield vis (Amira and Osirix will do this)
- time dependent data
- quantitative evaluation of the data
- large data (Microview will do 4GB)
- Mockup of UI
- key differntiating features
- Review software
- Osirix: Allen
- based on VTk/ITK
- more for clinicians and scientists
- strength is viewing of data before reading
- thumbnails from contents of databases
- management of databases is nice
- does support multiple, indendent windows
- volume rendering based on isosurface, with some preset values for skin, bone,
- uses control/status bar at the top of display
- mouse control is flexible, can be reassigned by user
- quick scrolling through images
- some segmentation support-but did not work as expected
- intensity/contrast setting apply to all views at once
- 2D view is strong, 3D is usable
- to take from this
- flexible view combinations, configurable
- each subwindow should have its own context
- envision a window like the map3d browser to manage the display and database
- in each pane, have a way of setting features of the display in that pane
- have a menu of color maps that the user can drop onto specific panes; think about color table management
- include a watermark generator in saving images section
- questions
- what sort of operations do we plan to support?, e.g., image processing
- granularity of parameters will have to evolve; start with coarse groupings of settings and spit out only as needed
- Microview: David (windows)
- used tab views of interface elements, but user has to manually delete tabs; tabs hold the provenance and allow user to reset a previous setting, much like a module in SCIRun
- had slider bar for brightness/contrast, worked on all panes
- lacked opacity settings for isosurface display, not clear how to get multiple surface
- no undo button for setting transfer function
- good help function, pop up
- good labeling, appears to be user settable
- color scheme was poor overall
- transfer function had histogram based interface that was fairly complex and hard to use
- UI is not great, nonintuitive even though layouts sometimes good
- keeps some history and show the values changed in a separate window
- isosurface render was very fast
- to take from this:
- nice layout
- consider how to maintain provenance of operation
- question:
- could overlay geometry that it read from a file over the rendering --it this useful?
- how do we manage provenance information
- scriptability: use Python? Need to figure this out
- ITK Snap: James
- has interesting interface to filters
- offers preview function
- provides step by step instructions on segmentation process
- segmentation algo did not actually work
- needed more reset functions
- to take from this
- preview is nice function when setting parameters
- has tutorial button at start up
- frame questions in terms of intuitive settings, e.g., higher quality, lower quality, or slower/faster, rather than asking to set numbers that may not mean anything
- questions
- Amira
- Review of collaborator needs
- Dave
- Jeroen
- Rob
- Candidate Data sets
- Jeroen will ask for multimodel cell data
- torso data (CT and MRI): ue those from Eric Voth
- head dataset from Fanxiang
- this looks good
- Osirix: Allen
- Next week
- UI: Joeren
- Image processing features: Allen
- Rendering options and features: James
- Scripting: David
