ActsPlugins::SeedClassifier Class Reference

Onnx model implementation for seed scoring and selection. More...

#include </home/runner/work/acts/acts/Plugins/Onnx/include/ActsPlugins/Onnx/SeedClassifier.hpp>

Public Member Functions
	SeedClassifier (const char *modelPath)
	Construct the scoring algorithm.
std::vector< std::vector< float > >	inferScores (NetworkBatchInput &networkInput) const
	Compute a score for each seed to be used in the seed selection.
std::vector< std::size_t >	seedSelection (std::vector< std::vector< std::size_t > > &clusters, std::vector< std::vector< float > > &outputTensor, float minSeedScore=0.1) const
	Select the seed associated with each cluster based on the score vector.
std::vector< std::size_t >	solveAmbiguity (std::vector< std::vector< std::size_t > > &clusters, NetworkBatchInput &networkInput, float minSeedScore=0.1) const
	Select the seed associated with each cluster.

Detailed Description

Onnx model implementation for seed scoring and selection.

Constructor & Destructor Documentation

SeedClassifier()

ActsPlugins::SeedClassifier::SeedClassifier

(

const char *

modelPath

)

Construct the scoring algorithm.

Parameters

modelPath

path to the model file

Member Function Documentation

inferScores()

std::vector< std::vector< float > > ActsPlugins::SeedClassifier::inferScores

(

NetworkBatchInput &

networkInput

)

const

Compute a score for each seed to be used in the seed selection.

Parameters

networkInput

input of the network

Returns

a vector of vector of seed score. Due to the architecture of the network each seed only have a size 1 score vector.

seedSelection()

std::vector< std::size_t > ActsPlugins::SeedClassifier::seedSelection	(	std::vector< std::vector< std::size_t > > &	clusters,
		std::vector< std::vector< float > > &	outputTensor,
		float	minSeedScore = 0.1 ) const

Select the seed associated with each cluster based on the score vector.

Parameters

clusters	is a vector of clusters, each cluster corresponds to a vector of seedIDs
outputTensor	is the score vector obtained from inferScores.
minSeedScore	is the minimum score a seed needs to be selected

Returns

a vector of seedIDs corresponding tho the good seeds

solveAmbiguity()

std::vector< std::size_t > ActsPlugins::SeedClassifier::solveAmbiguity	(	std::vector< std::vector< std::size_t > > &	clusters,
		NetworkBatchInput &	networkInput,
		float	minSeedScore = 0.1 ) const

Select the seed associated with each cluster.

Parameters

clusters	is a map of clusters, each cluster correspond to a vector of seed ID
networkInput	input of the network
minSeedScore	is the minimum score a seed need to be selected

Returns

a vector of seedID corresponding the the good seeds