Skip to content

Conversation

@SachidanandAlle
Copy link
Collaborator

@SachidanandAlle SachidanandAlle commented Nov 22, 2024

More details: #1785

Things fixed/added.

  • Auto Segmentation (Segmentation, DeepEdit, VISTA3D)
  • Interactive/Point Segmentation (Deepgrow, DeepEdit, SAM2D/SAM3D, VISTA3D)
  • Class Prompts (Select fewer classes for segmentation models and update the labels accordingly)
  • Support Configs/Options overrides
  • Support single slice update in case of Deepgrow 2D and SAM 2D models
  • Fix MONAI Label mode in study list
  • Add Active Learning Tab (Option to select next sample, and trigger training)

@sedghi @tangy5 @diazandr3s

Signed-off-by: Sachidanand Alle <salle@nvidia.com>
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
SachidanandAlle and others added 2 commits November 22, 2024 03:04
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
@SachidanandAlle SachidanandAlle enabled auto-merge (squash) November 22, 2024 11:16
@sedghi
Copy link
Contributor

sedghi commented Nov 22, 2024

Oh did you fix the MONAI Label mode enable/disable? I will review. BTW it is 3.9, not 3.10

Copy link
Contributor

@sedghi sedghi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

see my comments, great stuff

@SachidanandAlle
Copy link
Collaborator Author

see my comments, great stuff

thanks for the review. will update few things based on the comments.
will prepare the rc2 version and if all good then, we will release the new main version early next week.

Signed-off-by: Sachidanand Alle <salle@nvidia.com>
@SachidanandAlle SachidanandAlle changed the title Upgrade OHIF v3 plugin to support OHIF 3.10 version Upgrade OHIF v3 plugin to support OHIF 3.9+ version Nov 23, 2024
Signed-off-by: Sachidanand Alle <salle@nvidia.com>
@SachidanandAlle SachidanandAlle merged commit da0247e into main Nov 23, 2024
@SachidanandAlle SachidanandAlle deleted the sachi/ohifv3 branch November 23, 2024 02:48
@Omnipotent-coder0
Copy link

Hey @SachidanandAlle, I wanted to use MONAI models for latest OHIF version (currently 3.11 beta),
I am creating a segmentation by doing something like this 👇
` const addSeg = async () => {
const { servicesManager } = props;
const { segmentationService } = servicesManager.services;

const { viewportGridService, displaySetService } = servicesManager.services;
const { viewports, activeViewportId } = viewportGridService.getState();
const viewport = viewports.get(activeViewportId);
const displaySet = displaySetService.getDisplaySetByUID(
  viewport.displaySetInstanceUIDs[0]
);
// return { viewport, displaySet };

// const displaySet = servicesManager.services.displaySetService.getDisplaySetByUID(viewport.displaySetInstanceUIDs[0]);
console.log({ displaySet });
const labels = ['Test Segment 1', 'Test Segment 2', 'Test Segment 3'];
const createSegmentationPayload = {
  segmentationId: '555',
  segments: labels.reduce((acc, label, index) => {
    acc[index + 1] = {
      label,
      active: index === 0, // First segment is active
      locked: false,
    };
    return acc;
  }, {}),
  label: "Test segmentation"
}

await segmentationService.createLabelmapForDisplaySet(displaySet, createSegmentationPayload);

segmentationService.addOrUpdateSegmentation(createSegmentationPayload);

await segmentationService.addSegmentationRepresentation(viewport.viewportId, {
  segmentationId: createSegmentationPayload.segmentationId,
});
console.log("hello");

}`

which works fine,
but I can't say same for the volumeObject which comes out to be null
file : plugins/ohifv3/extensions/monai-label/src/components/MonaiLabelPanel.tsx
line no. 336 : const volumeLoadObject = segmentationService.getLabelmapVolume('1');
I assume the reason could be that OHIF viewer now supports stack-based segmentation rather than volume-based segmentation,
It would be very helpful, if you could provide me some guidance for this.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants