"Timestamp","Live Wire","What could be improved for the above tool?","2D Fast Marching","What could be improved for the above tool?","Region Growing","What could be improved for the above tool?","Add, Subtract, Wipe, Erase","What could be improved for the above tool?","Paint, Fill","What could be improved for the above tool?","Threshold","What could be improved for the above tool?","UL Threshold","What could be improved for the above tool?","OTSU","What could be improved for the above tool?","Fast Marching 3D","What could be improved for the above tool?","Region Growing 3D","What could be improved for the above tool?","Watershed","What could be improved for the above tool?","Picking","What could be improved for the above tool?","Slice Interpolation 2D/3D","What could be improved for the above tool?","Segmentation Utilities: Boolean Operations","What could be improved for the above tool?","Segmentation Utilities: Contour to Image","What could be improved for the above tool?","Segmentation Utilities: Image Masking","What could be improved for the above tool?","Segmentation Utilities: Morphological Operations","What could be improved for the above tool?","Segmentation Utilities: Surface to Image","What could be improved for the above tool?","Segmentation Utilities: Convert to MultiLabel","What could be improved for the above tool?","Free text space:","Regarding documentation?","Regarding user experience and user interface? (e.g. interaction patterns, short cuts, on screen text/feedback)?","Regarding offered functionality of existing tools?","Regarding new tools/algorithm? (e.g. integration of new methods or specific trained DL algorithms)","Do you use any other applications to mitigate problems with MITK segmentation tools? Which application are you using? What is great about it?","What is your segmentation use case / goal? Please provide us a short description","What is your overall satisfaction score with MITK segmentation?","Any general feedback or comment you would like to give and that helps us to improve MITK?"
"2021/06/04 10:48:16 am EET","I don't know it.","","I don't know it.","","I know it, but don't use it.","","OK :| but should be improved.","","Positive :) keep it as is!","","I don't know it.","","I don't know it.","","I don't know it.","","I don't know it.","","I don't know it.","","I don't know it.","","I don't know it.","","","","","","","","","","","","","","","","","Documentation is ok, but could be improved, especially the part that focuses on developers that will use it.","Good and usable, could become a little more minimal, especially some plugins.","Pretty good.","Very good, maybe some times the documentation is lacking.","","Lesion segmentation, multilabel segmentation.","4",""
"2021/06/07 9:06:53 am EET","I know it, but don't use it.","","I don't know it.","","Negative :( want to use it, but doesn't work out for me.","","OK :| but should be improved.","Erase should work independent of selected segmentation class.","Positive :) keep it as is!","","OK :| but should be improved.","Adaptive / Depth dependent would be great. ","I don't know it.","","I don't know it.","","I don't know it.","","I don't know it.","","I don't know it.","","I don't know it.","","I know it, but don't use it.","","I don't know it.","","I don't know it.","","I know it, but don't use it.","","I don't know it.","","I don't know it.","","I don't know it.","","I didn't work with the latest release (but the version before).
It should be a warning if presets are loaded twice, or a preset is loaded on top of a already definded segmentation class.
Furthermore batch processing could be improved (create segmentation with preset X for all images). Additionally it would be great if one could fix this preset, so that clinical partners can not accedently add more labels.
All in all, the tool works fine. These would just be some nice to have features.","","","","","","Segment and classify ultrasound data.","2","Clinical partners would be happy about a IOS (tablet) version"
"2021/06/07 9:49:44 am EET","OK :| but should be improved.","- expose parametrization so that the user can select how the wire should behave. Preset does not always work well
- 2d slice interpolation does not work for the created slices while the tools is active. One needs to disable and then reenable live wire for that to work","Negative :( want to use it, but doesn't work out for me.","- documentation is absent. What do the parameters mean?","OK :| but should be improved.","- Nice tool, but needs its parameters exposed! Defaults don't work all the time","Positive :) keep it as is!","","Positive :) keep it as is!","","OK :| but should be improved.","threshold tools need unification, see detailed answer on next page","OK :| but should be improved.","threshold tools need unification, see detailed answer on next page","OK :| but should be improved.","threshold tools need unification, see detailed answer on next page","Negative :( want to use it, but doesn't work out for me.","- documentation! What are the parameters doing?
- Can take a while to compute, so it is annoying that the algorithm is rerun with every seed one sets. We should be able to disable automatic execution so that we can set multiple seeds and then re-trigger it","Negative :( want to use it, but doesn't work out for me.","- buggy tool that does not do what I need it to do. Very weird behavior overall. Come see me (Fabian Isensee) for detailed feedback (ref for myself: la_005)","OK :| but should be improved.","Documentation is strange: ""Remark: The tool is (due to its implementation) faster if you use lower levels. So in doubt you should start with high levels (supposed to undersegment) and then lower the levels until you are happy."" This makes no sense - why start with the slow execution?","Negative :( want to use it, but doesn't work out for me.","- this tool loses a lot of its value by restricting the selection to a single region
- change this tool so that you can select multiple regions (shift click should toggle selection)
- after confirming, the active segmentation on the segmentation plugin should be set to the newly created one (currently this is the old one that includes all regions, not just the selected one)","OK :| but should be improved.","Fantastic tool overall. I have open issues on phabricator on this tool:
- needs multiprocessing (2d interp slices can be computed independently form each other)
- does not work for very large images
No issue about this yet:
- confirm all slices-> axial/coronal/sagittal are in a different order all the time. Please make it consistent.
Apart from those it's perfect and one of the best features MITK has to offer","OK :| but should be improved.","- Can be rather slow on large images. This is an elementwise bit operation, so it should be really fast. I don't know why this would be so slow
- multilabel images not supported? :-(","I don't know it.","","I know it, but don't use it.","","I know it, but don't use it.","","I don't know it.","","Negative :( want to use it, but doesn't work out for me.","- why does this want an image as input? Shouldn't that be a segmentation?","The thresholding tools need unification. They all do the same, just the way the thresholds are defined differs. So instead of having them each as a separate tool, just make a single thresholding tools through which all the functionality is exposed. Also why is this a 3D tool? This can also be used on 2D images and the fact that it is listed under 3d is confusing. Maybe just create a separate 'image-wide tools' section?","- Documentation is missing important information about the parametrization of some techniques (see detailed responses page1)
- hovering over icons with the mouse should display a short description of what each tool does (+ usage instructions like 'shift click to add seed points')","- hovering over icons with the mouse should display a short description of what each tool does (+ usage instructions like 'shift click to add seed points')","","- ilastik-like active learning with ML algorithm under the hood would be extremely useful, but probably hard to implement. We should first focus on the existing tools
- deep-learning-based active learning would be even cooler","I find myself often exporting images to nifti, loading them in python, running needed operations there and exporting them back to mitk haha. This is mostly to work around bugs in MITK such as problems with large images + 2d slice interpolation","I am heading the HIP ACVL and we do everything :-) I have 100+ segmentation datasets. Whatever you need for testing, I probably have it. Small, large, MR, MR, cryoET, ultrasound, anisotropic, isotropic, you name it! Please reach out (Fabian Isensee)!","4","I love you guys (and gals) <3"