Sean Doyle

I'm sorry - wasn't sure of of some of the steps and let this slip. I'll take a look tonight. If we follow @hackermd's recommendation then we should also change...

@hackermd Gladly!

One thing I found confusing about the API is the structure of the array of ImageLibraryEntry objects. It's an array of an array of `ImageLibraryEntry`. Initially I had specified an...

@hackermd - Yes - this works in the sense that it passes the pixelmed validator. Thanks! @fedorov - I agree that it looks like TID 1601 isn't required and that...

@hackermd Sure - I'll give it a try tonight. I'm finding this very odd - the reason I want to use the image library descriptors is so I can store...

> > The constructor of `ImageLibrary` could then accept `Sequence[Tuple[ImageLibraryEntryDescriptors, Sequence[ImageLibraryEntry]]]`, where each group item may contain 1 `ImageLibraryEntryDescriptors` instance (for the collection or group of images) and 1-n `ImageLibraryEntry`...

Thanks - I agree in general with discouraging free text but the app that I'm creating permits radiologists to enter text (this is their requirement) and don't want to generate...

@hackermd - do you mean like I did in the first message on this thread? The problem (perhaps a small implementation problem) is that in when a value with a...

If we did move the B-values to a standard location - would new SOPInstances need to be created? I'm guessing that the old private values would be left in place...

@CPBridge - I agree with continuing to allow segmentations with no 'positive' pixels. As we create pipelines of models that can pass inference results this makes sense. But I do...