KJB
 All Classes Namespaces Files Functions Variables Typedefs Enumerations Enumerator Friends Macros Groups Pages
Namespaces | Typedefs | Functions
tracking_metrics.h File Reference
#include <m_cpp/m_matrix.h>
#include <tracking_cpp/tracking_entity.h>
#include <tracking_cpp/tracking_trajectory.h>
#include <boost/bimap.hpp>
#include <vector>

Go to the source code of this file.

Namespaces

 kjb
 Classes and functions for dealing with trajectory files.
 
 kjb::tracking
 

Typedefs

typedef boost::bimap
< Entity_id, Entity_id > 
kjb::tracking::Correspondence
 

Functions

std::vector< Correspondencekjb::tracking::get_correspondence (const Canonical_trajectory_map &gt_trajs, const Canonical_trajectory_map &hyp_trajs, double threshold=1.0)
 Compute the correspondence between the groud truth track and hypothesized tracks. More...
 
void kjb::tracking::get_mt_ml_fragment_and_id_switch (const std::vector< Correspondence > &corrs, const Canonical_trajectory_map &gt_trajs, const Canonical_trajectory_map &hyp_trajs, double &mt, double &ml, size_t &frags, size_t &id_switch)
 Compute most_tracked, most_lost and id_switches for a given correspondence. More...
 
void kjb::tracking::get_mota_and_motp (const std::vector< size_t > &mme_ct, const std::vector< size_t > &fp_ct, const std::vector< size_t > &miss_ct, const std::vector< size_t > &match_ct, const std::vector< size_t > &obj_ct, const std::vector< double > &dists, double &mota, double &motp)
 Compute metrics based on the counts. More...
 
void kjb::tracking::init_ids_and_points (std::vector< Entity_id > &ids, std::vector< const Vector * > &points, const Canonical_trajectory_map &trajs, size_t cur_frame, const Correspondence &existing_corr, bool ground_truth)
 Extract unclaimed points for this time slice (ignores anything currently in exsiting correspondence) More...
 
void kjb::tracking::get_pw_distance (const std::vector< const Vector * > &pts1, const std::vector< const Vector * > &pts2, Matrix &distance)
 Compute the pairwise distance between pts1 and pts2 and stores the results in Matrix distance. More...
 
void kjb::tracking::get_best_matching (const kjb::Matrix &pw_distance, double threshold, std::vector< std::pair< int, int > > &matching)
 Compute the best matching using hungarian algorithm. More...
 
void kjb::tracking::get_counts_and_distances (const std::vector< Correspondence > &corrs, const Canonical_trajectory_map &gt_trajs, const Canonical_trajectory_map &hyp_trajs, std::vector< size_t > &mme_ct, std::vector< size_t > &fp_ct, std::vector< size_t > &miss_ct, std::vector< size_t > &match_ct, std::vector< size_t > &obj_ct, std::vector< double > &dists)
 Compute the counts (match, miss match, false positives, grouth truth objects, and the distance between matched grouth truth and hypothesized tracks. More...