opentl::core::cvdata::T_MEAS_FEAT Class Reference

Structure holding a single feature-level measurement. A single measurement is related to: 1 target, 1 modality, 1 camera, 1 thread. More...

List of all members.

Public Member Functions

std::size_t getFeatElementDim ()
 Get size/length of a single feature element.
std::size_t getJacSize ()
 Get degree of freedom of H matrix.
std::size_t getNMeas ()
 Get number of allocated features.
bool getUseMHypos ()
 Return true, if this measurement contains multi-hypotheses data structures for z and E.
void resizeAll (std::size_t nMeas_, std::size_t featElementDim_, std::size_t jacSize_=0, std::vector< std::size_t > *multiHypos=NULL)
 Resizes all vectors within structure and sets internal variables accordingly.
void resizeAssoc (std::size_t nMeas_, std::size_t featElementDim_)
 Resizes association vector within structure and sets internal variables accordingly.
void resizeE (std::size_t nMeas_, std::size_t featElementDim_, std::vector< std::size_t > *multiHypos=NULL)
 Resizes E within structure and sets internal variables accordingly.
void resizeH (std::size_t nMeas_, std::size_t featElementDim_)
 Resizes h within structure and sets internal variables accordingly.
void resizeJac (std::size_t nMeas_, std::size_t featElementDim_, std::size_t jacSize_)
 Resizes Jacobian of h within structure and sets internal variables accordingly.
void resizeR (std::size_t nMeas_, std::size_t featElementDim_)
 Resizes invCov within structure and sets internal variables accordingly.
void resizeZ (std::size_t nMeas_, std::size_t featElementDim_, std::vector< std::size_t > *multiHypos=NULL)
 Resizes z within structure and sets internal variables accordingly.
void setNMeas (std::size_t nMeas_)
 Set number of available features (it may be less than the number of allocated size!).
 T_MEAS_FEAT (std::size_t nMeas_=1, std::size_t featElementDim_=0, std::size_t jacSize_=0, std::vector< std::size_t > *multiHypos=NULL)

Public Attributes

double alpha
 alpha = missing detection rate
std::vector< int > assoc
 assoc = vector of flags, for associated data
std::vector< std::vector
< double > > 		E_mHypo
 E_mHypo/_sHypo = Residuals in projected feature space (z,h) with multiple/single hypotheses Dimensions:
  • outer (1st) nMeas: respective feature index ( = nMeas*featElementDim + intra element index)
  • inner (2nd) nMeas: (multi) measurement hypothesis index. The 2nd nMeas has to be resized at filling time, since its size can be different for every feature block.

std::vector< double > E_sHypo
std::vector< double > h
 h = vector of expected image features
std::vector< double > invCov
 invCov = invert covariances for E. ATTENTION: invCov encodes a covariance matrix (featElementDim * featElementDim) in the 1st dim. Index:
  • respective feature index ( = nMeas*featElementDim*featElementDim + intra matrix index)

std::vector< std::vector
< double > > 		Jac
 Jac = Jacobians of h. See jacSize parameter in resizeAll(). Dimensions:
  • outer (1st) nMeas: state dimension
  • inner (2nd) nMeas: respective feature index.

double lambda
 lambda = false alarms rate
std::vector< std::vector
< double > > 		z_mHypo
 Number of missing detections for this measurement (expected h, but no measurement observed).
std::vector< double > z_sHypo

Detailed Description

Structure holding a single feature-level measurement. A single measurement is related to: 1 target, 1 modality, 1 camera, 1 thread.

Constructor & Destructor Documentation

opentl::core::cvdata::T_MEAS_FEAT::T_MEAS_FEAT ( std::size_t  nMeas_ = 1,
std::size_t  featElementDim_ = 0,
std::size_t  jacSize_ = 0,
std::vector< std::size_t > *  multiHypos = NULL 
) [inline]

Member Function Documentation

std::size_t opentl::core::cvdata::T_MEAS_FEAT::getFeatElementDim (  )  [inline]

Get size/length of a single feature element.

std::size_t opentl::core::cvdata::T_MEAS_FEAT::getJacSize (  )  [inline]

Get degree of freedom of H matrix.

std::size_t opentl::core::cvdata::T_MEAS_FEAT::getNMeas (  )  [inline]

Get number of allocated features.

bool opentl::core::cvdata::T_MEAS_FEAT::getUseMHypos (  )  [inline]

Return true, if this measurement contains multi-hypotheses data structures for z and E.

void opentl::core::cvdata::T_MEAS_FEAT::resizeAll ( std::size_t  nMeas_,
std::size_t  featElementDim_,
std::size_t  jacSize_ = 0,
std::vector< std::size_t > *  multiHypos = NULL 
) [inline]

Resizes all vectors within structure and sets internal variables accordingly.

Parameters:
nMeas_ Number of features.
featElementDim_ Length of one feature element.
jacSize_ 1st nMeas of matrix H

void opentl::core::cvdata::T_MEAS_FEAT::resizeAssoc ( std::size_t  nMeas_,
std::size_t  featElementDim_ 
) [inline]

Resizes association vector within structure and sets internal variables accordingly.

Parameters:
nMeas_ Number of features.
featElementDim_ Length of one feature element.

void opentl::core::cvdata::T_MEAS_FEAT::resizeE ( std::size_t  nMeas_,
std::size_t  featElementDim_,
std::vector< std::size_t > *  multiHypos = NULL 
) [inline]

Resizes E within structure and sets internal variables accordingly.

Parameters:
nMeas_ Number of features.
featElementDim_ Length of one feature element.

void opentl::core::cvdata::T_MEAS_FEAT::resizeH ( std::size_t  nMeas_,
std::size_t  featElementDim_ 
) [inline]

Resizes h within structure and sets internal variables accordingly.

Parameters:
nMeas_ Number of features.
featElementDim_ Length of one feature element.

void opentl::core::cvdata::T_MEAS_FEAT::resizeJac ( std::size_t  nMeas_,
std::size_t  featElementDim_,
std::size_t  jacSize_ 
) [inline]

Resizes Jacobian of h within structure and sets internal variables accordingly.

Parameters:
nMeas_ Number of features.
featElementDim_ Length of one feature element.
jacSize_ 1st nMeas of matrix H.

void opentl::core::cvdata::T_MEAS_FEAT::resizeR ( std::size_t  nMeas_,
std::size_t  featElementDim_ 
) [inline]

Resizes invCov within structure and sets internal variables accordingly.

Parameters:
nMeas_ Number of features.
featElementDim_ Length of one feature element.

void opentl::core::cvdata::T_MEAS_FEAT::resizeZ ( std::size_t  nMeas_,
std::size_t  featElementDim_,
std::vector< std::size_t > *  multiHypos = NULL 
) [inline]

Resizes z within structure and sets internal variables accordingly.

Parameters:
nMeas_ Number of features.
featElementDim_ Length of one feature element.

void opentl::core::cvdata::T_MEAS_FEAT::setNMeas ( std::size_t  nMeas_  )  [inline]

Set number of available features (it may be less than the number of allocated size!).

Member Data Documentation

double opentl::core::cvdata::T_MEAS_FEAT::alpha

alpha = missing detection rate

std::vector<int> opentl::core::cvdata::T_MEAS_FEAT::assoc

assoc = vector of flags, for associated data

std::vector<std::vector<double> > opentl::core::cvdata::T_MEAS_FEAT::E_mHypo

E_mHypo/_sHypo = Residuals in projected feature space (z,h) with multiple/single hypotheses Dimensions:

  • outer (1st) nMeas: respective feature index ( = nMeas*featElementDim + intra element index)
  • inner (2nd) nMeas: (multi) measurement hypothesis index. The 2nd nMeas has to be resized at filling time, since its size can be different for every feature block.

SINGLE measurement hypothesis (2nd nMeas size = 1): example usage with a featElementDim of 3 (e.g for RGB) 2nd nMeas -> 1st E_sHypo[0]=feat0_e_R d E_sHypo[1]=feat0_e_G i E_sHypo[2]=feat0_e_B m E_sHypo[3]=feat1_e_R E_sHypo[4]=feat1_e_G E_sHypo[5]=feat1_e_G ...

MULTIPLE measurement hypothesis: example usage with a featElementDim of 3 (e.g for RGB) 2nd nMeas -> 1st E_mHypo[0][0]=feat0_mHypo0_e_R E_mHypo[0][1]=feat0_mHypo1_e_R d E_mHypo[1][0]=feat0_mHypo0_e_G E_mHypo[1][1]=feat0_mHypo1_e_G i E_mHypo[2][0]=feat0_mHypo0_e_B E_mHypo[2][1]=feat0_mHypo1_e_B ... m E_mHypo[3][0]=feat1_mHypo0_e_R E_mHypo[3][1]=feat1_mHypo1_e_R E_mHypo[4][0]=feat1_mHypo0_e_G E_mHypo[4][1]=feat1_mHypo1_e_G E_mHypo[5][0]=feat1_mHypo0_e_G E_mHypo[5][1]=feat1_mHypo1_e_G ...

std::vector<double> opentl::core::cvdata::T_MEAS_FEAT::E_sHypo

std::vector<double> opentl::core::cvdata::T_MEAS_FEAT::h

h = vector of expected image features

std::vector<double> opentl::core::cvdata::T_MEAS_FEAT::invCov

invCov = invert covariances for E. ATTENTION: invCov encodes a covariance matrix (featElementDim * featElementDim) in the 1st dim. Index:

  • respective feature index ( = nMeas*featElementDim*featElementDim + intra matrix index)

Cov. matrix invCov for a SINGLE measurement hypothesis (2nd dim size = 1): example usage with a featElementDim of 3 (e.g for RGB) 2nd dim -> 1st R[0]=feat0_R*R d R[1]=feat0_R*G i R[2]=feat0_R*B m R[3]=feat0_G*R R[4]=feat0_G*G R[5]=feat0_G*B R[6]=feat0_B*R R[7]=feat0_B*G R[8]=feat0_B*B R[9]=feat1_R*R R[10]=feat1_R*G ...

std::vector<std::vector<double> > opentl::core::cvdata::T_MEAS_FEAT::Jac

Jac = Jacobians of h. See jacSize parameter in resizeAll(). Dimensions:

  • outer (1st) nMeas: state dimension
  • inner (2nd) nMeas: respective feature index.

Example: if your pose dim = 6 and you have 100 features in h, H will be of dim 6 x 100

double opentl::core::cvdata::T_MEAS_FEAT::lambda

lambda = false alarms rate

std::vector<std::vector<double> > opentl::core::cvdata::T_MEAS_FEAT::z_mHypo

Number of missing detections for this measurement (expected h, but no measurement observed).

z = vector of image features associated to each h (with possibly multiple-hypo)

std::vector<double> opentl::core::cvdata::T_MEAS_FEAT::z_sHypo

Generated on Fri Jul 31 17:38:38 2009 for OpenTL by  doxygen 1.5.8