You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As stated elsewhere, we do not really need the Pillow pixel data handler anymore, now that pygdcm is easily installable, and we have the pylibjpeg plugins.
I would go even further and say that we may solely rely on the pylibjpeg plugins as soon as they implement all needed transfer syntaxes (e.g. not less than GDCM).
What we need for this is a functioning libjpeg-turbo plugin as an alternative to the GPL-licensed libjpeg plugin, and if it does not implement JPEG-LS (I'm not sure), convert pylibjpegls to a pylibjpeg plugin, otherwise retire it.
The advantage is that these are in our control (don't have to hunt bugs in Pillow and GDCM) and we have a unified handling of all plugins.
The disadvantage is that these are in our control (e.g. we have to maintain them). There may also be some advantage (feature or performance-wise) in Pillow or GDCM that I don't see at the moment (the handling of in-memory datasets in GDCM?)
This is at the moment more of a discussion issue, but I wanted to hear other opinions on this...
The text was updated successfully, but these errors were encountered:
...we do not really need the Pillow pixel data handler anymore
...we may solely rely on the pylibjpeg plugins
I think that is reasonable.
What we need for this is a functioning libjpeg-turbo plugin as an alternative to the GPL-licensed libjpeg plugin
Well, the build of the library is working, and PyTurboJPEG can find it, so I just need to get around to properly connecting the draft decode/decode_pixel_data functions.
There may also be some advantage (feature or performance-wise) in Pillow or GDCM that I don't see at the moment (the handling of in-memory datasets in GDCM?)
From a technical viewpoint, I don't see any reason the plugins would be prevented from using something in the underyling libraries, so they could be adapted to new features, if we have the time.
As stated elsewhere, we do not really need the Pillow pixel data handler anymore, now that
pygdcm
is easily installable, and we have thepylibjpeg
plugins.I would go even further and say that we may solely rely on the
pylibjpeg
plugins as soon as they implement all needed transfer syntaxes (e.g. not less than GDCM).What we need for this is a functioning
libjpeg-turbo
plugin as an alternative to the GPL-licensedlibjpeg
plugin, and if it does not implement JPEG-LS (I'm not sure), convertpylibjpegls
to apylibjpeg
plugin, otherwise retire it.The advantage is that these are in our control (don't have to hunt bugs in Pillow and GDCM) and we have a unified handling of all plugins.
The disadvantage is that these are in our control (e.g. we have to maintain them). There may also be some advantage (feature or performance-wise) in Pillow or GDCM that I don't see at the moment (the handling of in-memory datasets in GDCM?)
This is at the moment more of a discussion issue, but I wanted to hear other opinions on this...
The text was updated successfully, but these errors were encountered: