PYME.recipes.machine_learning module

class PYME.recipes.machine_learning.CNNFilter(parent=None, **kwargs)

Bases: PYME.recipes.base.Filter

Use a previously trained Keras neural network to filter the data. Used for learnt de-noising and/or deconvolution. Runs prediction piecewise over the image with over-lapping ROIs and averages the prediction results.

Notes

Keras and either Tensorflow or Theano must be installed set up for this module to work. This is not a default dependency of python-microscopy as the conda-installable versions don’t have GPU support.

Attributes

default_view
hide_in_overview
inputs
outputs
pipeline_view
pipeline_view_min

Methods

applyFilter(data, chanNum, frNum, im)
completeMetadata(im)
trait_items_event(event_trait,name,items_event)
trait_property_changed(...)
traits_init()
traits_inited([True])
applyFilter(data, chanNum, frNum, im)
completeMetadata(im)
class PYME.recipes.machine_learning.svmSegment(parent=None, **kwargs)

Bases: PYME.recipes.base.Filter

Attributes

default_view
hide_in_overview
inputs
outputs
pipeline_view
pipeline_view_min

Methods

applyFilter(data, chanNum, frNum, im)
completeMetadata(im)
trait_items_event(event_trait,name,items_event)
trait_property_changed(...)
traits_init()
traits_inited([True])
applyFilter(data, chanNum, frNum, im)
completeMetadata(im)