Post-Registration Evaluation#
Find the best list of transformations given multiple lists and the images they were registered from
Parameters#
- image_stacks
Image stacks as a dictionary of numpy arrays or list of HDF5 dataset URI’s
- transformations
The resulting transformations for each of the image_stacks
- url
Url of an hdf5 Dataset to save the chosen transformed image stack
- reference_stack
Most relevant stack name
Outputs#
- image_stacks
Image stacks as a dictionary of numpy arrays or list of HDF5 dataset URI’s
- transformations
The resulting transformations for each of the image_stacks
- reference_stack
Most relevant stack name
- ranked_stack_names
Stack names ordered from most to least relevant
Reg2DPostEvaluation#
Given several stacks of images and their image transformations, determine the stack and list of transformations that results in the best registration.
- Identifier:
ewoksndreg.tasks.reg2d_posteval.Reg2DPostEvaluation- Task type:
class
- Required inputs:
image_stacks, transformations
- Optional inputs:
output_configuration, output_root_uri, reference_stack, skip
- Outputs:
image_stacks, output_configuration, ranked_stack_names, reference_stack, transformations
Notes#
Deciding whether an alignment is “good” is very hard to solve algorithmically, therefore the attemps to find the best alignment in this task are in no way definitive.
- For every image_stacks there are two measures used:
Mean squared error between the images
Smoothness of the transformations using change of corner coordinates when applying transformations