Rework Slicer segmentation plugin

NVIDIA / ai-assisted-annotation-client

Client side integration example source code and libraries for AI-Assisted Annotation SDK

Other

308 stars 64 forks source link

Rework Slicer segmentation plugin #47

Closed lassoan closed 4 years ago

lassoan commented 4 years ago

Reworked GUI layout
Fixed volume data caching
Add options to disable compression of input image to improve speed on fast local network connections
Show warning before uploading data to remote server
Save last 10 server addresses
Added 3 step-by-step tutorials - see here: https://github.com/lassoan/ai-assisted-annotation-client/tree/slicer-ui-rework/slicer-plugin
Removed support for Slicer-4.10.2, as SimpleITK bundled in this Slicer version did not seem to be compatible with latest version of AIAA client library (this is not a big issue, as we'll release new Slicer Stable Release soon and until then the Preview Releases can be used)

SachidanandAlle commented 4 years ago

Merged.. but i see a small issue.. when i do auto-segment (say for spleen) it creates new segment instead of editing/modifying the selected one.... any reasons for that?

Screenshot from 2019-12-16 11-12-17

lassoan commented 4 years ago

Auto-segment may create several segments from scratch, without any user inputs. It is for the user's convenience that he does not have to pre-create any segments but all the segments are created automatically.

Annotation is different because it requires user input, which the user may want to tune and then re-run the annotation. Also, most annotation models create only a single segment.

You can find example workflow descriptions here: https://github.com/NVIDIA/ai-assisted-annotation-client/tree/master/slicer-plugin#tutorials-and-examples. Let me know if you have suggestions for improvements.

SachidanandAlle commented 4 years ago

Auto-segment may create several segments from scratch, without any user inputs. It is for the user's convenience that he does not have to pre-create any segments but all the segments are created automatically.

Annotation is different because it requires user input, which the user may want to tune and then re-run the annotation. Also, most annotation models create only a single segment.

You can find example workflow descriptions here: https://github.com/NVIDIA/ai-assisted-annotation-client/tree/master/slicer-plugin#tutorials-and-examples. Let me know if you have suggestions for improvements.

i see now.. we have changed the first behavior here to make easier for user.. :)

after loading the image, user need not to create any label for auto-segmentation... user can directly select the model and run auto-segmentation.. 👍