Python export image to numpy is always float32
It would be really nice to preserve the output data type, because right now, if we try to convert a big 8bit image to numpy, this will come out 4x bigger and since we need a copy in case the app ref is lost, it requires x8 times the image footprint of available RAM...
This does not change anything to set the app's output parameter pixel type.
Ideally this could also apply to the python copy of the array.
Is it something feasible without a lot of code or do you think this should go in the 9.0 roadmap ?
Edited by Vincent Delbar