Package org.anchoranalysis.image.inference.segment

Class SegmentedObjects

Object

SegmentedObjects

public class SegmentedObjects
extends Object

Objects that are a result of an instance-segmentation.

Each object has an associated confidence score, and an associated class-label.

Internally, objects are partitioned by class-label.

Author:

Owen Feehan

Constructor Summary

Constructors

Constructor	Description
SegmentedObjects(String classLabel, Collection<WithConfidence<MultiScaleObject>> objects, SegmentedBackground background, ExecutionTimeRecorder executionTimeRecorder)	Create for a collection of objects with an identical label.
SegmentedObjects(List<LabelledWithConfidence<MultiScaleObject>> objects, SegmentedBackground background, ExecutionTimeRecorder executionTimeRecorder)	Create for a collection of objects with potentially differing labels.

Method Summary

Modifier and Type	Method	Description
DualScale<SegmentedObjectsAtScale>	getObjects()	The segmented-objects, at two different scales.
Optional<WithConfidence<MultiScaleObject>>	highestConfidence()	The object-mask with the highest confidence.
boolean	isEmpty()	Whether no segmented objects exist.
SegmentedObjects	reduce(ReduceElements<ObjectMask> reduce, boolean separateEachLabel, ExecutionTimeRecorder executionTimeRecorder)	Reduces the segmented-objects, applying a reduction algorithm separately to each object-class.
int	size()	The total number of segmented objects.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

SegmentedObjects

public SegmentedObjects(String classLabel,
 Collection<WithConfidence<MultiScaleObject>> objects,
 SegmentedBackground background,
 ExecutionTimeRecorder executionTimeRecorder)

Create for a collection of objects with an identical label.

Parameters:

classLabel - the label that applies to each object in objects.

objects - the objects with label classLabel.

background - background-images used for visualizing the segmentation, at two respective scales.

executionTimeRecorder - records the execution-time of particular operations.

SegmentedObjects

public SegmentedObjects(List<LabelledWithConfidence<MultiScaleObject>> objects,
 SegmentedBackground background,
 ExecutionTimeRecorder executionTimeRecorder)

Create for a collection of objects with potentially differing labels.

Parameters:

objects - the objects that are the result of the segmentation, with associated confidence and labels.

background - background-images used for visualizing the segmentation, at two respective scales.

executionTimeRecorder - records the execution-time of particular operations.

Method Details

reduce

public SegmentedObjects reduce(ReduceElements<ObjectMask> reduce,
 boolean separateEachLabel,
 ExecutionTimeRecorder executionTimeRecorder)
                        throws OperationFailedException

Reduces the segmented-objects, applying a reduction algorithm separately to each object-class.

Parameters:

reduce - the algorithm used to reduce each object-class.

separateEachLabel - if true, each label is reduced separately. if false, all labels are reduced together.

executionTimeRecorder - records the execution-time of particular operations.

Returns:

a new SegmentedObjects with the reduce algorithm applied to each object-class, reusing the existing objects.

Throws:

OperationFailedException - if the reduction fails on any object-class.

highestConfidence

public Optional<WithConfidence<MultiScaleObject>> highestConfidence()

The object-mask with the highest confidence.

Returns:

the highest-confidence object-mask or Optional.empty() if no objects exist.

size

public int size()

The total number of segmented objects.

Returns:

the total number.

isEmpty

public boolean isEmpty()

Whether no segmented objects exist.

Returns:

true if no segmented objects exist, false otherwise.

getObjects

public DualScale<SegmentedObjectsAtScale> getObjects()

The segmented-objects, at two different scales.