Applications and Limitations of Measurements and Segmentations

Support Development
1

Describe Your Question

Is there any high level documentation outlining the designs of the measurement and segmentation features? What they will do and what they won’t do by design? Roadmaps of future capabilities?

Like many others, I am looking at the OHIF Viewer as the basis of a web based annotation tool.

Playing with some of the OHIF/Orthanc docker demos, for example, it appears that measurement are tied to the specific series/instance on which they are drawn. They are not mapped into the other series at the same spatial locations. From the documentation, I don’t have a sense as to whether this would be possible or not.

The segmentations feature does appear to map segmentations into any series covering the same spatial location. Segmentations, however, currently lack the more finely controlled Freehand and Spline ROIs of the basic viewer measurement feature. The segmentation panel also lacks the Image Slice Sync feature that the basic viewer has to keep multiple contrasts aligned during scrolling.

Before I dive into code to try to figure this out, I thought I would ask if there is some high level documentation somewhere outlining the design goals, what’s been accomplished and what the current limitations are.

Thanks,
John.

2

Rodmap is here

https://ohif.org/roadmap/Q2-2024

3

@alireza
Could you provide the documentation that mentions “The segmentations feature does appear to map segmentations into any series covering the same spatial location.”

4

To clarify, I discovered this in practice by performing manual segmentation with a multi-contrast/series study. Regions I segmented in one contrast were automatically segmented (the labels were mapped) in the other contrasts in the same spatial locations. I could see my segmentation show up in all windows.

Were you asking for documentation of that? Screenshots? Or did I give the impression I had read about the feature in the documentation?

In our use case, we have multiple research projects that take a multi-contrast approach to identifying disease and therefore tend to utilize all contrasts while segmenting. This can partially be executed using the current segmentation menu, but the segmentation tools were limited when I last checked in May to non-irregular features like ellipses, rectangles or brushes.

Unfortunately, the diseases we want to segment are not regular shapes. The measurement tools do handle irregular shapes - but do not map to all contrasts. A measurement polygon in one contrast is not automatically drawn in all other contrasts.

John.