Presence of THUMBNAIL may result in incorrect placement of annotations

[fedorov]

This is about the same slide as in #305, and the slide jump is the same, but the new "twist" on the issue is that while zooming the annotations are showing up in the wrong place. There are no errors or warnings in the console. I tried this in Incognito, so should not be a caching issue.

https://viewers-sandbox-gha-testing.web.app/studies/1.2.826.0.1.3680043.8.498.98417242628600706149560301686974206083/series/1.2.826.0.1.3680043.8.498.84253944677190903991260180165776664353

@DanielaSchacherer says that Slim recognizes only 6 out of 7 instances in the corresponding series, with OVERVIEW instance not being recognized. But even if there is an issue with the data, Slim should not show annotations in the wrong place!

[fedorov]

Another slide with the same problem, also missing OVERVIEW per Daniela.

https://viewers-sandbox-gha-testing.web.app/studies/1.2.826.0.1.3680043.8.498.63505735290979224089200866845268211842/series/1.2.826.0.1.3680043.8.498.8182925743242083930807624952710331478

Also same problem here: https://viewers-sandbox-gha-testing.web.app/studies/1.2.826.0.1.3680043.8.498.29801694699643154195584396190079068396/series/1.2.826.0.1.3680043.8.498.43620887593209104223243153615387310193

[fedorov]

@DanielaSchacherer for this one, I do see errors in the log once I zoom in to the base layer (I think).

https://viewers-sandbox-gha-testing.web.app/studies/1.2.826.0.1.3680043.8.498.29801694699643154195584396190079068396/series/1.2.826.0.1.3680043.8.498.43620887593209104223243153615387310193

[DanielaSchacherer]

@DanielaSchacherer for this one, I do see errors in the log once I zoom in to the base layer (I think).

I can confirm that.

[DanielaSchacherer]

Hi @igoroctaviano,

sorry to bring up other things in this issue, but I think this could be related and Andrey already mentioned this above:

• OVERVIEW image is not listed in the DICOM Tag Browser of Slim - is that by purpose? The overview image is shown in the left side pannel though.
• For three of the five samples, I can not fully zoom out anymore (see image below). I think I could do that when I checked last week, but I am not 100% sure.

Do you remember in which commits you did bigger changes and to which commits I could go back to do some testing on that problems?

[DanielaSchacherer]

This is about the same slide as in #305, and the slide jump is the same, but the new "twist" on the issue is that while zooming the annotations are showing up in the wrong place. There are no errors or warnings in the console. I tried this in Incognito, so should not be a caching issue.

https://viewers-sandbox-gha-testing.web.app/studies/1.2.826.0.1.3680043.8.498.98417242628600706149560301686974206083/series/1.2.826.0.1.3680043.8.498.84253944677190903991260180165776664353

@DanielaSchacherer says that Slim recognizes only 6 out of 7 instances in the corresponding series, with OVERVIEW instance not being recognized. But even if there is an issue with the data, Slim should not show annotations in the wrong place!

Just wanted to add that I checked the slide in QuPath - no problems occured here.

[igoroctaviano]

Hi @igoroctaviano,

sorry to bring up other things in this issue, but I think this could be related and Andrey already mentioned this above:

• OVERVIEW image is not listed in the DICOM Tag Browser of Slim - is that by purpose? The overview image is shown in the left side pannel though.
• For three of the five samples, I can not fully zoom out anymore (see image below). I think I could do that when I checked last week, but I am not 100% sure.

Do you remember in which commits you did bigger changes and to which commits I could go back to do some testing on that problems?

Do you think this is a new issue, or something you’ve just never noticed before?

Regarding the zoom behavior: it only works differently when you’re not using a thumbnail image — which only happens if there’s an equivalent volume image available.
If you do have a thumbnail but no volume image, then zoom behaves just like it did before.

Maybe you can join today’s call so we can all sync up — that way we can better understand if these are new issues or regressions.

[fedorov]

@erikogabrielsson for your awareness - we don't know what is going on, but somehow the presence of wsidicomizer-generated thumbnail image breaks Slim - related to imi-bigpicture/wsidicomizer#134, linking for awareness.

[fedorov]

For the dataset in question, we decided to proceed with releasing the dataset omitting thumbnail images. We will need to get back to this eventually.

[igoroctaviano]

@fedorov do we need to move this one to backlog and label it differently?