feat(adapters): overlapping segs with labelmap images

[pedrokohler]

Context

In order to address OHIF/Viewers#3496 we're taking a two steps approach:

1 - chopping the volume data buffer array into planar data buffer arrays
2 - adopt some strategy to duplicate these planar data buffer arrays when we get overlapping segs

This PR aims to cover the second step.

Changes & Results

Before - loss of information:

After - overlapping segs maintained:

Testing

Checklist

PR

• [] My Pull Request title is descriptive, accurate and follows the
  semantic-release format and guidelines.

Code

• [] My code has been well-documented (function documentation, inline comments,
  etc.)

• [] I have run the yarn build:update-api to update the API documentation, and have
  committed the changes to this PR. (Read more here https://www.cornerstonejs.org/docs/contribute/update-api)

Public Documentation Updates

• [] The documentation page has been updated as necessary for any public API
  additions or removals.

Tested Environment

• [] "OS:
• [] "Node version:
• [] "Browser:

[netlify]

✅ Deploy Preview for cornerstone-3d-docs ready!

Name	Link
🔨 Latest commit	4b6a761
🔍 Latest deploy log	https://app.netlify.com/sites/cornerstone-3d-docs/deploys/67a3e0e15b6e270008bc08c0
😎 Deploy Preview	https://deploy-preview-1815--cornerstone-3d-docs.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[sedghi]

I don't think we're on the right track here. The adapter should return imageIds, and the example/consumer/ohif should create a single segmentation from those imageIds. We might need to tweak the internal segmentation addition and hydration process to make this work, but it's pretty likely to be okay. For now, let's put the volume one on the back burner - we'll need more optimizations to create sub-volumes for each imageId group, and that's a bit more complicated since we'd need to do it only where they're actually located. So make sure the example work with having one segmentation, the segmentation is consisted of multiple imageIds

[pedrokohler]

I don't think we're on the right track here. The adapter should return imageIds, and the example/consumer/ohif should create a single segmentation from those imageIds. We might need to tweak the internal segmentation addition and hydration process to make this work, but it's pretty likely to be okay. For now, let's put the volume one on the back burner - we'll need more optimizations to create sub-volumes for each imageId group, and that's a bit more complicated since we'd need to do it only where they're actually located. So make sure the example work with having one segmentation, the segmentation is consisted of multiple imageIds

I was really hoping you would point to specific parts of the code and ask for specific changes instead of abstract ones. I'm not following what you're asking me to do.

[sedghi]

here we should have the segmentationId, there is one segmentation for this case and it happens to be overlapping, so don't change this to an array

[sedghi]

same here, there is only one segmentation here

[sedghi]

So basically the idea is to have this work.

const { labelMapImages } = 
        await Cornerstone3D.Segmentation.createFromDICOMSegBuffer(
            referenceImageIds,
            arrayBuffer,
            {
                metadataProvider: metaData,
                skipOverlapping: false
            }
        );

Then we do

csToolsSegmentation.addSegmentations([
        {
            segmentationId,
            representation: {
                type: cornerstoneTools.Enums.SegmentationRepresentations
                    .Labelmap,
                data: {
                    imageIds: labelMapImages.flat()
                }
            }
        }
    ]);

The key files that need to be changed are

https://github.com/cornerstonejs/cornerstone3D/blob/main/packages/tools/src/eventListeners/segmentation/imageChangeEventListener.ts

Here

const derivedImageId = getCurrentLabelmapImageIdForViewport(
        viewportId,
        representation.segmentationId
      );

as you see it only get one labelmapimageId for viewport => but we have changed that so that it can have multiple for one segmentationId (overlapping)

and then later on in that file

viewport.addImages([
        {
          imageId: derivedImageId,
          representationUID: `${segmentationId}-${SegmentationRepresentations.Labelmap}`,
          callback: ({ imageActor }) => {
            imageActor.getMapper().setInputData(imageData);
          },
        },
      ]);

you see we only add one derivedImageId to the viewport (since we only had support for one slice -> one imageId -> per labelmap) but now we want to support multiple

so we need to loop over derivedImageIds for that slice and then add them one by one which should work

Hope this helps