Markus D. Herrmann

@scaramallion I took a shot at the proposed approach with https://github.com/MGHComputationalPathology/highdicom/pull/36 (see [highdicom.io.ImageFileReader._decode_frame()](https://github.com/MGHComputationalPathology/highdicom/blob/2cf248304fee1cea6b16af4d1f47f0b0bab74987/src/highdicom/io.py#L553). To get this to work, I had to re-implement quite a bit of the functionality of the...

@scaramallion I ultimately went down a slightly different path (see #1447). Instead of implementing a separate `decode_frame()` method for each pixel handler, I found a way to decode individual frames...

@scaramallion @darcymason any thoughts on this PR? I would prefer including this functionality in *pydicom* rather than keeping it in *highdicom*, since I consider it generally useful beyond the creation...

@darcymason sorry I somehow completely missed your comments on the PR. Will start addressing them.

See also #1243

That would be fine with me, too. Either way, I think it would be important that the ICC Profile is applied by default upon access of `pixel_array`.

> I'm not fond of applying stuff to the pixel data without explicit instructions from users. Decoders need to consider other attributes as well (e.g., `PhotometricInterpretation` and `PixelPresentation`), and I...

In the meantime, we implemented color management in highdicom (https://github.com/MGHComputationalPathology/highdicom/blob/master/src/highdicom/color.py)

@scaramallion one of the challenges in implementing this feature in pydicom has been the fact that the numpy and Pillow packages are optional dependencies. This is unfortunate, since it makes...

@AlVinci we are currently implementing a high-level interface for reading frames of a multi-frame Part10 file (see [highdicom PR](https://github.com/MGHComputationalPathology/highdicom/pull/36)). We hope to get all the low-level functions ultimately into the...