Steve Pieper

Hi Emel - I'm not following your question exactly. The way we've been handling things is that the SEG contains the pixel data of the segmentation (and things like anatomic...

> writing dicom sr for segmentations/referring them. does OHIF support that? I'm pretty sure nobody has done that yet. What I would do is look at the structure of the...

As we work on this can we add some [tests for the sr code](https://github.com/dcmjs-org/dcmjs/blob/c8fba616e53e36db4cd80f7c334580490898c4f7/test/test_sr.js) just to show how all the pieces should fit together?

Thanks for the report 👍 Could you provide a PR with a test that replicates this issue? @wayfarer3130 can you have a look?

That sounds like a regression, probably introduced by [this commit](https://github.com/dcmjs-org/dcmjs/commit/70b24332783d63c9db2ed21d512d9f7b526c5222) or similar. Calling `naturalizeDataset` should be idempotent in my opinion, especially if it was that way before. It would be...

That looks like a codec problem (the circle of noise). This probably causes the incorrect pixel values.

Thanks for working on this 👍 Let's be sure to give the final PR commit line a semantic versioning prefix like `feat(typescript)` so it'll auto deploy to npm.

Yes, Danny set it up and I used it for the first time the other day: https://github.com/dcmjs-org/dcmjs/issues/159#issuecomment-767717261

@WoonchanCho this looks great! 💯 It'll be a huge help and will encourage us to add more comments.

Yes, makes sense. Can you add a link to an example dataset (maybe put it on dcmjs-org/data).