gtsam 4.1.1
gtsam
gtsam::MFAS Class Reference

Public Member Functions

 MFAS (const std::map< KeyPair, double > &edgeWeights)
 Construct from the weighted directed edges between the nodes. More...
 
 MFAS (const TranslationEdges &relativeTranslations, const Unit3 &projectionDirection)
 Constructor to be used in the context of translation averaging. More...
 
KeyVector computeOrdering () const
 Computes the 1D MFAS ordering of nodes in the graph. More...
 
std::map< KeyPair, double > computeOutlierWeights () const
 Computes the outlier weights of the graph. More...
 

Public Types

using KeyPair = std::pair< Key, Key >
 
using TranslationEdges = std::vector< BinaryMeasurement< Unit3 > >
 

Constructor & Destructor Documentation

◆ MFAS() [1/2]

gtsam::MFAS::MFAS ( const std::map< KeyPair, double > &  edgeWeights)
inline

Construct from the weighted directed edges between the nodes.

Each node is identified by a Key.

Parameters
edgeWeightsweights of edges in the graph

◆ MFAS() [2/2]

MFAS::MFAS ( const TranslationEdges &  relativeTranslations,
const Unit3 projectionDirection 
)

Constructor to be used in the context of translation averaging.

Here, the nodes of the graph are cameras in 3D and the edges have a unit translation direction between them. The weights of the edges is computed by projecting them along a projection direction.

Parameters
relativeTranslationstranslation directions between the cameras
projectionDirectiondirection in which edges are to be projected

Member Function Documentation

◆ computeOrdering()

KeyVector MFAS::computeOrdering ( ) const

Computes the 1D MFAS ordering of nodes in the graph.

Returns
orderedNodes: vector of nodes in the obtained order

◆ computeOutlierWeights()

map< MFAS::KeyPair, double > MFAS::computeOutlierWeights ( ) const

Computes the outlier weights of the graph.

We define the outlier weight of a edge to be zero if the edge is an inlier and the magnitude of its edgeWeight if it is an outlier. This function internally calls computeOrdering and uses the obtained ordering to identify outlier edges.

Returns
outlierWeights: map from an edge to its outlier weight.

The documentation for this class was generated from the following files: