Spine Plugin

Plugin for machine learning based localization, classification and segmentation of the spine. More...

Collaboration diagram for Spine Plugin:

Detailed Description

Plugin for machine learning based localization, classification and segmentation of the spine.

Classes
class	SpinePlugin
	Plugin class responsible for initializing all spine related components. More...
class	SpineBaseAlgorithm
	Spine localisation / classification and segmentation algorithm. More...
class	SpineBaseController
	Controller for Spine Algorithms. More...
class	SpinePluginFactory
	Factory for the spine workflow algorithms. More...
class	SpineControllerFactory
	Factory for the spine workflow controllers. More...
class	SpineDataAnnotationFactory
	Factory for the spine workflow controllers. More...
class	OrientedVertebra
	A representation of a vertebra, containing orientation, classification keypoints, segmentations, meshes and all other related information. More...
class	SpineData
	Container for set of vertebrae and other spine-related metadata The SpineData class is the main Data type of the ImFusion SpinePlugin. More...
class	VertebraDissectionAlgorithm
	Algorithm for extracting structures of interest from an instance of SpineData The structures are extracted by performing a deformable mesh registration of a template structure onto the given input SpineData. More...

Functions
std::vector< int64_t > SPINE_PLUGIN_API	hungarianAssignment (Eigen::MatrixXf const &costs, double nullCost, bool minimize=true)
	Calculates the optimal assignment for a matching problem.

Function Documentation

hungarianAssignment()

std::vector< int64_t > SPINE_PLUGIN_API hungarianAssignment	(	Eigen::MatrixXf const &	costs,
		double	nullCost,
		bool	minimize = true )

#include <SpinePlugin/include/ImFusion/Spine/HungarianMatching.h>

Calculates the optimal assignment for a matching problem.

Provide a matrix of N x N that states the costs of matching item col to item row If the matrix is not square, the nullCost parameter should be set to a positive value and is used as a cost for not assigning an item. The assignments are returned as a list of assignments, where -1 denotes a 'not assigned' state (again only if the matrix is not square)