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
For example ZeroMeanUnitVarianceTransformation only permits a single mean and standard deviation. However the spec here suggests these these should be arrays/lists, and an axes value should be supported as well. This means that it's not currently possible to specify using mean/std on a per-channel basis.
Update relevant methods (e.g. public Number getMean(); public void setMean(Number mean);) to work with either arrays or lists, rather than individual numbers, and also add axes support - but this will be an API-breaking change
Deprecate the current methods, and add new ones - but this would require different naming (e.g. getMeans()) that could deviate from other implementations
The text was updated successfully, but these errors were encountered:
We can absolutely convert the Transformations in a way to handle lists. For the IO classes in anything <=0.4.0 we just have to make sure that we transfer from a PRIMITIVE to List and vice versa so that there is only a single implementation of each transformation that handles a list.
In <= v0.4.0 some information is missing from classes in the io.bioimage.specification.transformation.transformation package.
For example
ZeroMeanUnitVarianceTransformation
only permits a single mean and standard deviation. However the spec here suggests these these should be arrays/lists, and anaxes
value should be supported as well. This means that it's not currently possible to specify using mean/std on a per-channel basis.ScaleLinearTransformation
(a possible alternative) andScaleMinMaxTransformation
have similar limitations.I can think of two main ways to fix it:
public Number getMean(); public void setMean(Number mean);
) to work with either arrays or lists, rather than individual numbers, and also add axes support - but this will be an API-breaking changegetMeans()
) that could deviate from other implementationsThe text was updated successfully, but these errors were encountered: