Point Cloud Library (PCL): pcl::MinCutSegmentation< PointT

This class implements the segmentation algorithm based on minimal cut of the graph. More...

#include <pcl/segmentation/min_cut_segmentation.h>

Inheritance diagram for pcl::MinCutSegmentation< PointT >:

Public Types
using	KdTree = pcl::search::Search< PointT >

using	KdTreePtr = typename KdTree::Ptr

using	PointCloud = pcl::PointCloud< PointT >

using	PointCloudConstPtr = typename PointCloud::ConstPtr

using	Traits = boost::adjacency_list_traits< boost::vecS, boost::vecS, boost::directedS >

using	mGraph = boost::adjacency_list< boost::vecS, boost::vecS, boost::directedS, boost::property< boost::vertex_name_t, std::string, boost::property< boost::vertex_index_t, long, boost::property< boost::vertex_color_t, boost::default_color_type, boost::property< boost::vertex_distance_t, long, boost::property< boost::vertex_predecessor_t, Traits::edge_descriptor > > > > >, boost::property< boost::edge_capacity_t, double, boost::property< boost::edge_residual_capacity_t, double, boost::property< boost::edge_reverse_t, Traits::edge_descriptor > > > >

using	CapacityMap = boost::property_map< mGraph, boost::edge_capacity_t >::type

using	ReverseEdgeMap = boost::property_map< mGraph, boost::edge_reverse_t >::type

using	VertexDescriptor = Traits::vertex_descriptor

using	EdgeDescriptor = boost::graph_traits< mGraph >::edge_descriptor

using	OutEdgeIterator = boost::graph_traits< mGraph >::out_edge_iterator

using	VertexIterator = boost::graph_traits< mGraph >::vertex_iterator

using	ResidualCapacityMap = boost::property_map< mGraph, boost::edge_residual_capacity_t >::type

using	IndexMap = boost::property_map< mGraph, boost::vertex_index_t >::type

using	InEdgeIterator = boost::graph_traits< mGraph >::in_edge_iterator

using	mGraphPtr = shared_ptr< mGraph >

Public Types inherited from pcl::PCLBase< PointT >
using	PointCloud = pcl::PointCloud< PointT >

using	PointCloudPtr = typename PointCloud::Ptr

using	PointCloudConstPtr = typename PointCloud::ConstPtr

using	PointIndicesPtr = PointIndices::Ptr

using	PointIndicesConstPtr = PointIndices::ConstPtr

Public Member Functions
	MinCutSegmentation ()
	Constructor that sets default values for member variables. More...

	~MinCutSegmentation ()
	Destructor that frees memory. More...

void	setInputCloud (const PointCloudConstPtr &cloud) override
	This method simply sets the input point cloud. More...

double	getSigma () const
	Returns normalization value for binary potentials. More...

void	setSigma (double sigma)
	Allows to set the normalization value for the binary potentials as described in the article. More...

double	getRadius () const
	Returns radius to the background. More...

void	setRadius (double radius)
	Allows to set the radius to the background. More...

double	getSourceWeight () const
	Returns weight that every edge from the source point has. More...

void	setSourceWeight (double weight)
	Allows to set weight for source edges. More...

KdTreePtr	getSearchMethod () const
	Returns search method that is used for finding KNN. More...

void	setSearchMethod (const KdTreePtr &tree)
	Allows to set search method for finding KNN. More...

unsigned int	getNumberOfNeighbours () const
	Returns the number of neighbours to find. More...

void	setNumberOfNeighbours (unsigned int neighbour_number)
	Allows to set the number of neighbours to find. More...

std::vector< PointT, Eigen::aligned_allocator< PointT > >	getForegroundPoints () const
	Returns the points that must belong to foreground. More...

void	setForegroundPoints (typename pcl::PointCloud< PointT >::Ptr foreground_points)
	Allows to specify points which are known to be the points of the object. More...

std::vector< PointT, Eigen::aligned_allocator< PointT > >	getBackgroundPoints () const
	Returns the points that must belong to background. More...

void	setBackgroundPoints (typename pcl::PointCloud< PointT >::Ptr background_points)
	Allows to specify points which are known to be the points of the background. More...

void	extract (std::vector< pcl::PointIndices > &clusters)
	This method launches the segmentation algorithm and returns the clusters that were obtained during the segmentation. More...

double	getMaxFlow () const
	Returns that flow value that was calculated during the segmentation. More...

mGraphPtr	getGraph () const
	Returns the graph that was build for finding the minimum cut. More...

pcl::PointCloud< pcl::PointXYZRGB >::Ptr	getColoredCloud ()
	Returns the colored cloud. More...

Public Member Functions inherited from pcl::PCLBase< PointT >
	PCLBase ()
	Empty constructor. More...

	PCLBase (const PCLBase &base)
	Copy constructor. More...

virtual	~PCLBase ()=default
	Destructor. More...

const PointCloudConstPtr	getInputCloud () const
	Get a pointer to the input point cloud dataset. More...

virtual void	setIndices (const IndicesPtr &indices)
	Provide a pointer to the vector of indices that represents the input data. More...

virtual void	setIndices (const IndicesConstPtr &indices)
	Provide a pointer to the vector of indices that represents the input data. More...

virtual void	setIndices (const PointIndicesConstPtr &indices)
	Provide a pointer to the vector of indices that represents the input data. More...

virtual void	setIndices (std::size_t row_start, std::size_t col_start, std::size_t nb_rows, std::size_t nb_cols)
	Set the indices for the points laying within an interest region of the point cloud. More...

IndicesPtr	getIndices ()
	Get a pointer to the vector of indices used. More...

const IndicesConstPtr	getIndices () const
	Get a pointer to the vector of indices used. More...

const PointT &	operator[] (std::size_t pos) const
	Override PointCloud operator[] to shorten code. More...

Protected Member Functions
bool	buildGraph ()
	This method simply builds the graph that will be used during the segmentation. More...

void	calculateUnaryPotential (int point, double &source_weight, double &sink_weight) const
	Returns unary potential(data cost) for the given point index. More...

bool	addEdge (int source, int target, double weight)
	This method simply adds the edge from the source point to the target point with a given weight. More...

double	calculateBinaryPotential (int source, int target) const
	Returns the binary potential(smooth cost) for the given indices of points. More...

bool	recalculateUnaryPotentials ()
	This method recalculates unary potentials(data cost) if some changes were made, instead of creating new graph. More...

bool	recalculateBinaryPotentials ()
	This method recalculates binary potentials(smooth cost) if some changes were made, instead of creating new graph. More...

void	assembleLabels (ResidualCapacityMap &residual_capacity)
	This method analyzes the residual network and assigns a label to every point in the cloud. More...

Protected Member Functions inherited from pcl::PCLBase< PointT >
bool	initCompute ()
	This method should get called before starting the actual computation. More...

bool	deinitCompute ()
	This method should get called after finishing the actual computation. More...

Protected Attributes
double	inverse_sigma_
	Stores the sigma coefficient. More...

bool	binary_potentials_are_valid_
	Signalizes if the binary potentials are valid. More...

double	epsilon_
	Used for comparison of the floating point numbers. More...

double	radius_
	Stores the distance to the background. More...

bool	unary_potentials_are_valid_
	Signalizes if the unary potentials are valid. More...

double	source_weight_
	Stores the weight for every edge that comes from source point. More...

KdTreePtr	search_
	Stores the search method that will be used for finding K nearest neighbors. More...

unsigned int	number_of_neighbours_
	Stores the number of neighbors to find. More...

bool	graph_is_valid_
	Signalizes if the graph is valid. More...

std::vector< PointT, Eigen::aligned_allocator< PointT > >	foreground_points_
	Stores the points that are known to be in the foreground. More...

std::vector< PointT, Eigen::aligned_allocator< PointT > >	background_points_
	Stores the points that are known to be in the background. More...

std::vector< pcl::PointIndices >	clusters_
	After the segmentation it will contain the segments. More...

mGraphPtr	graph_
	Stores the graph for finding the maximum flow. More...

std::shared_ptr< CapacityMap >	capacity_
	Stores the capacity of every edge in the graph. More...

std::shared_ptr< ReverseEdgeMap >	reverse_edges_
	Stores reverse edges for every edge in the graph. More...

std::vector< VertexDescriptor >	vertices_
	Stores the vertices of the graph. More...

std::vector< std::set< int > >	edge_marker_
	Stores the information about the edges that were added to the graph. More...

VertexDescriptor	source_
	Stores the vertex that serves as source. More...

VertexDescriptor	sink_
	Stores the vertex that serves as sink. More...

double	max_flow_
	Stores the maximum flow value that was calculated during the segmentation. More...

Protected Attributes inherited from pcl::PCLBase< PointT >
PointCloudConstPtr	input_
	The input point cloud dataset. More...

IndicesPtr	indices_
	A pointer to the vector of point indices to use. More...

bool	use_indices_
	Set to true if point indices are used. More...

bool	fake_indices_
	If no set of indices are given, we construct a set of fake indices that mimic the input PointCloud. More...

Detailed Description

template<typename PointT>
class pcl::MinCutSegmentation< PointT >

This class implements the segmentation algorithm based on minimal cut of the graph.

The description can be found in the article: "Min-Cut Based Segmentation of Point Clouds"

Author: : Aleksey Golovinskiy and Thomas Funkhouser.

Definition at line 59 of file min_cut_segmentation.h.

Member Typedef Documentation

◆ CapacityMap

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::CapacityMap = boost::property_map< mGraph, boost::edge_capacity_t >::type

Definition at line 87 of file min_cut_segmentation.h.

◆ EdgeDescriptor

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::EdgeDescriptor = boost::graph_traits<mGraph>::edge_descriptor

Definition at line 93 of file min_cut_segmentation.h.

◆ IndexMap

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::IndexMap = boost::property_map< mGraph, boost::vertex_index_t >::type

Definition at line 101 of file min_cut_segmentation.h.

◆ InEdgeIterator

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::InEdgeIterator = boost::graph_traits<mGraph>::in_edge_iterator

Definition at line 103 of file min_cut_segmentation.h.

◆ KdTree

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::KdTree = pcl::search::Search<PointT>

Definition at line 63 of file min_cut_segmentation.h.

◆ KdTreePtr

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::KdTreePtr = typename KdTree::Ptr

Definition at line 64 of file min_cut_segmentation.h.

◆ mGraph

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::mGraph = boost::adjacency_list< boost::vecS, boost::vecS, boost::directedS, boost::property< boost::vertex_name_t, std::string, boost::property< boost::vertex_index_t, long, boost::property< boost::vertex_color_t, boost::default_color_type, boost::property< boost::vertex_distance_t, long, boost::property< boost::vertex_predecessor_t, Traits::edge_descriptor > > > > >, boost::property< boost::edge_capacity_t, double, boost::property< boost::edge_residual_capacity_t, double, boost::property< boost::edge_reverse_t, Traits::edge_descriptor > > > >

Definition at line 85 of file min_cut_segmentation.h.

◆ mGraphPtr

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::mGraphPtr = shared_ptr<mGraph>

Definition at line 105 of file min_cut_segmentation.h.

◆ OutEdgeIterator

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::OutEdgeIterator = boost::graph_traits<mGraph>::out_edge_iterator

Definition at line 95 of file min_cut_segmentation.h.

◆ PointCloud

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::PointCloud = pcl::PointCloud<PointT>

Definition at line 65 of file min_cut_segmentation.h.

◆ PointCloudConstPtr

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::PointCloudConstPtr = typename PointCloud::ConstPtr

Definition at line 66 of file min_cut_segmentation.h.

◆ ResidualCapacityMap

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::ResidualCapacityMap = boost::property_map< mGraph, boost::edge_residual_capacity_t >::type

Definition at line 99 of file min_cut_segmentation.h.

◆ ReverseEdgeMap

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::ReverseEdgeMap = boost::property_map< mGraph, boost::edge_reverse_t>::type

Definition at line 89 of file min_cut_segmentation.h.

◆ Traits

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::Traits = boost::adjacency_list_traits< boost::vecS, boost::vecS, boost::directedS >

Definition at line 75 of file min_cut_segmentation.h.

◆ VertexDescriptor

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::VertexDescriptor = Traits::vertex_descriptor

Definition at line 91 of file min_cut_segmentation.h.

◆ VertexIterator

template<typename PointT >

using pcl::MinCutSegmentation< PointT >::VertexIterator = boost::graph_traits<mGraph>::vertex_iterator

Definition at line 97 of file min_cut_segmentation.h.

Constructor & Destructor Documentation

◆ MinCutSegmentation()

template<typename PointT >

pcl::MinCutSegmentation< PointT >::MinCutSegmentation

Constructor that sets default values for member variables.

Definition at line 50 of file min_cut_segmentation.hpp.

◆ ~MinCutSegmentation()

template<typename PointT >

pcl::MinCutSegmentation< PointT >::~MinCutSegmentation

Destructor that frees memory.

Definition at line 73 of file min_cut_segmentation.hpp.

Member Function Documentation

◆ addEdge()

template<typename PointT >

bool pcl::MinCutSegmentation< PointT >::addEdge	(	int	source,
		int	target,
		double	weight
	)

protected

This method simply adds the edge from the source point to the target point with a given weight.

Parameters

[in]	source	index of the source point of the edge
[in]	target	index of the target point of the edge
[in]	weight	weight that will be assigned to the (source, target) edge

Definition at line 424 of file min_cut_segmentation.hpp.

◆ assembleLabels()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::assembleLabels ( ResidualCapacityMap & residual_capacity )

protected

This method analyzes the residual network and assigns a label to every point in the cloud.

Parameters

[in] residual_capacity residual network that was obtained during the segmentation

Definition at line 535 of file min_cut_segmentation.hpp.

References pcl::PointIndices::indices.

◆ buildGraph()

template<typename PointT >

bool pcl::MinCutSegmentation< PointT >::buildGraph

protected

This method simply builds the graph that will be used during the segmentation.

Definition at line 303 of file min_cut_segmentation.hpp.

◆ calculateBinaryPotential()

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::calculateBinaryPotential	(	int	source,
		int	target
	)		const

protected

Returns the binary potential(smooth cost) for the given indices of points.

In other words it returns weight that must be assigned to the edge from source to target point.

Parameters

[in]	source	index of the source point of the edge
[in]	target	index of the target point of the edge

Definition at line 450 of file min_cut_segmentation.hpp.

References pcl::geometry::distance().

◆ calculateUnaryPotential()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::calculateUnaryPotential	(	int	point,
		double &	source_weight,
		double &	sink_weight
	)		const

protected

Returns unary potential(data cost) for the given point index.

In other words it calculates weights for (source, point) and (point, sink) edges.

Parameters

[in]	point	index of the point for which weights will be calculated
[out]	source_weight	calculated weight for the (source, point) edge
[out]	sink_weight	calculated weight for the (point, sink) edge

Definition at line 368 of file min_cut_segmentation.hpp.

◆ extract()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::extract ( std::vector< pcl::PointIndices > & clusters )

This method launches the segmentation algorithm and returns the clusters that were obtained during the segmentation.

The indices of points that belong to the object will be stored in the cluster with index 1, other indices will be stored in the cluster with index 0.

Parameters

[out] clusters clusters that were obtained. Each cluster is an array of point indices.

Definition at line 218 of file min_cut_segmentation.hpp.

◆ getBackgroundPoints()

template<typename PointT >

std::vector< PointT, Eigen::aligned_allocator< PointT > > pcl::MinCutSegmentation< PointT >::getBackgroundPoints

Returns the points that must belong to background.

Definition at line 200 of file min_cut_segmentation.hpp.

◆ getColoredCloud()

template<typename PointT >

pcl::PointCloud< pcl::PointXYZRGB >::Ptr pcl::MinCutSegmentation< PointT >::getColoredCloud

Returns the colored cloud.

Points that belong to the object have the same color.

Definition at line 563 of file min_cut_segmentation.hpp.

References pcl::PointCloud< PointT >::height, pcl::PointCloud< PointT >::is_dense, pcl::PointCloud< PointT >::points, and pcl::PointCloud< PointT >::width.

◆ getForegroundPoints()

template<typename PointT >

std::vector< PointT, Eigen::aligned_allocator< PointT > > pcl::MinCutSegmentation< PointT >::getForegroundPoints

Returns the points that must belong to foreground.

Definition at line 182 of file min_cut_segmentation.hpp.

◆ getGraph()

template<typename PointT >

pcl::MinCutSegmentation< PointT >::mGraphPtr pcl::MinCutSegmentation< PointT >::getGraph

Returns the graph that was build for finding the minimum cut.

Definition at line 296 of file min_cut_segmentation.hpp.

◆ getMaxFlow()

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::getMaxFlow

Returns that flow value that was calculated during the segmentation.

Definition at line 289 of file min_cut_segmentation.hpp.

◆ getNumberOfNeighbours()

template<typename PointT >

unsigned int pcl::MinCutSegmentation< PointT >::getNumberOfNeighbours

Returns the number of neighbours to find.

Definition at line 162 of file min_cut_segmentation.hpp.

◆ getRadius()

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::getRadius

Returns radius to the background.

Definition at line 112 of file min_cut_segmentation.hpp.

◆ getSearchMethod()

template<typename PointT >

pcl::MinCutSegmentation< PointT >::KdTreePtr pcl::MinCutSegmentation< PointT >::getSearchMethod

Returns search method that is used for finding KNN.

The graph is build such way that it contains the edges that connect point and its KNN.

Definition at line 148 of file min_cut_segmentation.hpp.

◆ getSigma()

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::getSigma

Returns normalization value for binary potentials.

For more information see the article.

Definition at line 94 of file min_cut_segmentation.hpp.

◆ getSourceWeight()

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::getSourceWeight

Returns weight that every edge from the source point has.

Definition at line 130 of file min_cut_segmentation.hpp.

◆ recalculateBinaryPotentials()

template<typename PointT >

bool pcl::MinCutSegmentation< PointT >::recalculateBinaryPotentials

protected

This method recalculates binary potentials(smooth cost) if some changes were made, instead of creating new graph.

Definition at line 490 of file min_cut_segmentation.hpp.

◆ recalculateUnaryPotentials()

template<typename PointT >

bool pcl::MinCutSegmentation< PointT >::recalculateUnaryPotentials

protected

This method recalculates unary potentials(data cost) if some changes were made, instead of creating new graph.

Definition at line 465 of file min_cut_segmentation.hpp.

◆ setBackgroundPoints()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setBackgroundPoints ( typename pcl::PointCloud< PointT >::Ptr background_points )

Allows to specify points which are known to be the points of the background.

Parameters

[in] background_points point cloud that contains background points.

Definition at line 207 of file min_cut_segmentation.hpp.

References pcl::PointCloud< PointT >::cbegin(), and pcl::PointCloud< PointT >::cend().

◆ setForegroundPoints()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setForegroundPoints ( typename pcl::PointCloud< PointT >::Ptr foreground_points )

Allows to specify points which are known to be the points of the object.

Parameters

[in] foreground_points point cloud that contains foreground points. At least one point must be specified.

Definition at line 189 of file min_cut_segmentation.hpp.

References pcl::PointCloud< PointT >::cbegin(), and pcl::PointCloud< PointT >::cend().

◆ setInputCloud()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setInputCloud ( const PointCloudConstPtr & cloud )

overridevirtual

This method simply sets the input point cloud.

Parameters

[in] cloud the const boost shared pointer to a PointCloud

Reimplemented from pcl::PCLBase< PointT >.

Definition at line 84 of file min_cut_segmentation.hpp.

◆ setNumberOfNeighbours()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setNumberOfNeighbours ( unsigned int neighbour_number )

Allows to set the number of neighbours to find.

Parameters

[in] neighbour_number new number of neighbours

Definition at line 169 of file min_cut_segmentation.hpp.

◆ setRadius()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setRadius ( double radius )

Allows to set the radius to the background.

Parameters

[in] radius new radius to the background

Definition at line 119 of file min_cut_segmentation.hpp.

◆ setSearchMethod()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setSearchMethod ( const KdTreePtr & tree )

Allows to set search method for finding KNN.

The graph is build such way that it contains the edges that connect point and its KNN.

Parameters

[in] tree search method that will be used for finding KNN.

Definition at line 155 of file min_cut_segmentation.hpp.

◆ setSigma()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setSigma ( double sigma )

Allows to set the normalization value for the binary potentials as described in the article.

Parameters

[in] sigma new normalization value

Definition at line 101 of file min_cut_segmentation.hpp.

◆ setSourceWeight()

template<typename PointT >

void pcl::MinCutSegmentation< PointT >::setSourceWeight ( double weight )

Allows to set weight for source edges.

Every edge that comes from the source point will have that weight.

Parameters

[in] weight new weight

Definition at line 137 of file min_cut_segmentation.hpp.

Member Data Documentation

◆ background_points_

template<typename PointT >

std::vector<PointT, Eigen::aligned_allocator<PointT> > pcl::MinCutSegmentation< PointT >::background_points_

protected

Stores the points that are known to be in the background.

Definition at line 293 of file min_cut_segmentation.h.

◆ binary_potentials_are_valid_

template<typename PointT >

bool pcl::MinCutSegmentation< PointT >::binary_potentials_are_valid_

protected

Signalizes if the binary potentials are valid.

Definition at line 266 of file min_cut_segmentation.h.

◆ capacity_

template<typename PointT >

std::shared_ptr<CapacityMap> pcl::MinCutSegmentation< PointT >::capacity_

protected

Stores the capacity of every edge in the graph.

Definition at line 302 of file min_cut_segmentation.h.

◆ clusters_

template<typename PointT >

std::vector<pcl::PointIndices> pcl::MinCutSegmentation< PointT >::clusters_

protected

After the segmentation it will contain the segments.

Definition at line 296 of file min_cut_segmentation.h.

◆ edge_marker_

template<typename PointT >

std::vector< std::set<int> > pcl::MinCutSegmentation< PointT >::edge_marker_

protected

Stores the information about the edges that were added to the graph.

It is used to avoid the duplicate edges.

Definition at line 311 of file min_cut_segmentation.h.

◆ epsilon_

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::epsilon_

protected

Used for comparison of the floating point numbers.

Definition at line 269 of file min_cut_segmentation.h.

◆ foreground_points_

template<typename PointT >

std::vector<PointT, Eigen::aligned_allocator<PointT> > pcl::MinCutSegmentation< PointT >::foreground_points_

protected

Stores the points that are known to be in the foreground.

Definition at line 290 of file min_cut_segmentation.h.

◆ graph_

template<typename PointT >

mGraphPtr pcl::MinCutSegmentation< PointT >::graph_

protected

Stores the graph for finding the maximum flow.

Definition at line 299 of file min_cut_segmentation.h.

◆ graph_is_valid_

template<typename PointT >

bool pcl::MinCutSegmentation< PointT >::graph_is_valid_

protected

Signalizes if the graph is valid.

Definition at line 287 of file min_cut_segmentation.h.

◆ inverse_sigma_

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::inverse_sigma_

protected

Stores the sigma coefficient.

It is used for finding smooth costs. More information can be found in the article.

Definition at line 263 of file min_cut_segmentation.h.

◆ max_flow_

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::max_flow_

protected

Stores the maximum flow value that was calculated during the segmentation.

Definition at line 320 of file min_cut_segmentation.h.

◆ number_of_neighbours_

template<typename PointT >

unsigned int pcl::MinCutSegmentation< PointT >::number_of_neighbours_

protected

Stores the number of neighbors to find.

Definition at line 284 of file min_cut_segmentation.h.

◆ radius_

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::radius_

protected

Stores the distance to the background.

Definition at line 272 of file min_cut_segmentation.h.

◆ reverse_edges_

template<typename PointT >

std::shared_ptr<ReverseEdgeMap> pcl::MinCutSegmentation< PointT >::reverse_edges_

protected

Stores reverse edges for every edge in the graph.

Definition at line 305 of file min_cut_segmentation.h.

◆ search_

template<typename PointT >

KdTreePtr pcl::MinCutSegmentation< PointT >::search_

protected

Stores the search method that will be used for finding K nearest neighbors.

Neighbours are used for building the graph.

Definition at line 281 of file min_cut_segmentation.h.

◆ sink_

template<typename PointT >

VertexDescriptor pcl::MinCutSegmentation< PointT >::sink_

protected

Stores the vertex that serves as sink.

Definition at line 317 of file min_cut_segmentation.h.

◆ source_

template<typename PointT >

VertexDescriptor pcl::MinCutSegmentation< PointT >::source_

protected

Stores the vertex that serves as source.

Definition at line 314 of file min_cut_segmentation.h.

◆ source_weight_

template<typename PointT >

double pcl::MinCutSegmentation< PointT >::source_weight_

protected

Stores the weight for every edge that comes from source point.

Definition at line 278 of file min_cut_segmentation.h.

◆ unary_potentials_are_valid_

template<typename PointT >

bool pcl::MinCutSegmentation< PointT >::unary_potentials_are_valid_

protected

Signalizes if the unary potentials are valid.

Definition at line 275 of file min_cut_segmentation.h.

◆ vertices_

template<typename PointT >

std::vector< VertexDescriptor > pcl::MinCutSegmentation< PointT >::vertices_

protected

Stores the vertices of the graph.

Definition at line 308 of file min_cut_segmentation.h.

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

pcl/segmentation/min_cut_segmentation.h
pcl/segmentation/impl/min_cut_segmentation.hpp

Public Types

Public Member Functions

Protected Member Functions

Protected Attributes

Detailed Description

template<typename PointT> class pcl::MinCutSegmentation< PointT >

Member Typedef Documentation

◆ CapacityMap

◆ EdgeDescriptor

◆ IndexMap

◆ InEdgeIterator

◆ KdTree

◆ KdTreePtr

◆ mGraph

◆ mGraphPtr

◆ OutEdgeIterator

◆ PointCloud

◆ PointCloudConstPtr

◆ ResidualCapacityMap

◆ ReverseEdgeMap

◆ Traits

◆ VertexDescriptor

◆ VertexIterator

Constructor & Destructor Documentation

◆ MinCutSegmentation()

◆ ~MinCutSegmentation()

Member Function Documentation

◆ addEdge()

◆ assembleLabels()

◆ buildGraph()

◆ calculateBinaryPotential()

◆ calculateUnaryPotential()

◆ extract()

◆ getBackgroundPoints()

◆ getColoredCloud()

◆ getForegroundPoints()

◆ getGraph()

◆ getMaxFlow()

◆ getNumberOfNeighbours()

◆ getRadius()

◆ getSearchMethod()

◆ getSigma()

◆ getSourceWeight()

◆ recalculateBinaryPotentials()

◆ recalculateUnaryPotentials()

◆ setBackgroundPoints()

◆ setForegroundPoints()

◆ setInputCloud()

◆ setNumberOfNeighbours()

◆ setRadius()

◆ setSearchMethod()

◆ setSigma()

◆ setSourceWeight()

Member Data Documentation

◆ background_points_

◆ binary_potentials_are_valid_

◆ capacity_

◆ clusters_

◆ edge_marker_

◆ epsilon_

◆ foreground_points_

◆ graph_

◆ graph_is_valid_

◆ inverse_sigma_

◆ max_flow_

◆ number_of_neighbours_

◆ radius_

◆ reverse_edges_

◆ search_

◆ sink_

◆ source_

◆ source_weight_

◆ unary_potentials_are_valid_

◆ vertices_

template<typename PointT>
class pcl::MinCutSegmentation< PointT >