Markus D. Herrmann

I generally agree. My only concern is that there may not necessary be an easier way, but just different ways. And introducing another way to describe the segmentation may be...

What do you think about adding the`voi_lut_transformation: Optional[VOILUTTransformation]` parameter and including a deprecation warning for `window_width` and `window_center`? That would make it easier to deprecate the parameters in a future...

See also https://github.com/pydicom/pydicom/issues/1414

We should avoid the global configuration as much as possible. My preferred approach would be to add a parameter to the constructor of `Dataset` and friends (`Sequence`, etc.) such as...

> I would argue that we should probably throw a more useful error message and leave it up to the user to check that the segmentation is non-empty before trying...

> Actually @adamvonpaternos can you please confirm whether you are encountering this issue when there are no empty frames or all empty frames? Your text "if there are no empty...

Good point @pieper. The question arises whether such a segmentation can be encoded sparsely since this would result in a Pixel Data element with an empty value. The value of...

Agreed! We need to allow segments with all values being zero. Whether an "empty" segmentation could be encoded in a sparse fashion is something that probably would require a change...

@CPBridge I would suggest not raising an error when `omit_empty_frames` is `True` and no non-empty frames exist, but instead log a warning message and encode all frames. If one performs...

@fedorov Your points are well taken and I can see how an "empty" segment may cause confusion in certain situations. However, the same may be true for very small objects,...