boofcv.abst.feature.associate

Interface Associate

All Known Subinterfaces:

AssociateDescription<Desc>, AssociateDescription2D<Desc>, AssociateMaxDistance<D>

All Known Implementing Classes:

AssociateDescTo2D, AssociateMaxDistanceNaive, AssociateNearestNeighbor, BaseAssociateLocation2DFilter, EnforceUniqueByScore, EnforceUniqueByScore.Describe, EnforceUniqueByScore.Describe2D, WrapAssociateGreedy, WrapAssociateSurfBasic

public interface Associate

Common interface for associating features between two images. Found associations are returned in a list of AssociatedIndex which specifies the index and score of the matching pair. Implementing classes can optionally ensure that a unique pairing is found from source to destination and/or the reverse. See functions uniqueSource() and uniqueDestination(). Indexes refer to the index in the input list for source and destination lists. Inputs are not specified in this interface but are specified in a child interface.

DESIGN NOTES:
Indexes of matching features are used instead of the descriptions because descriptions are often separated from another more complex data structure and the index can be easily matched to that data.
Unassociated feature lists can be easily computed using the returned set of associations. This functionality is provided since in some cases it can be computed at virtually no cost during association.

Method Summary

Modifier and Type	Method and Description
void	associate() Finds the best match for each item in the source list with an item in the destination list.
org.ddogleg.struct.FastQueue<AssociatedIndex>	getMatches() List of associated features.
MatchScoreType	getScoreType() Specifies the type of score which is returned.
org.ddogleg.struct.GrowQueue_I32	getUnassociatedDestination() Indexes of features in the destination set which are not associated.
org.ddogleg.struct.GrowQueue_I32	getUnassociatedSource() Indexes of features in the source set which are not associated.
void	setThreshold(double score) Associations are only considered if their score is less than or equal to the specified threshold.
boolean	uniqueDestination() If at most one match is returned for each destination feature.
boolean	uniqueSource() If at most one match is returned for each source feature.

Method Detail

associate

void associate()

Finds the best match for each item in the source list with an item in the destination list.

getMatches

org.ddogleg.struct.FastQueue<AssociatedIndex> getMatches()

List of associated features. Indexes refer to the index inside the input lists.

Returns:

List of associated features.

getUnassociatedSource

org.ddogleg.struct.GrowQueue_I32 getUnassociatedSource()

Indexes of features in the source set which are not associated. WARNING: In some implementations the unassociated list is recomputed each time this function is invoked. In other implementations it was found virtually for free while the matches are found.

Returns:

List of unassociated source features by index.

getUnassociatedDestination

org.ddogleg.struct.GrowQueue_I32 getUnassociatedDestination()

Indexes of features in the destination set which are not associated. WARNING: In some implementations the unassociated list is recomputed each time this function is invoked. In other implementations it was found virtually for free while the matches are found.

Returns:

List of unassociated destination features by index.

setThreshold

void setThreshold(double score)

Associations are only considered if their score is less than or equal to the specified threshold. To remove any threshold test set this value to Double.MAX_VALUE

Parameters:

score - The threshold.

getScoreType

MatchScoreType getScoreType()

Specifies the type of score which is returned.

Returns:

Type of association score.

uniqueSource

boolean uniqueSource()

If at most one match is returned for each source feature.

Returns:

true for unique source association

uniqueDestination

boolean uniqueDestination()

If at most one match is returned for each destination feature.

Returns:

true for unique destination association