Class KPCluster

Class Documentation

class KPCluster

Stores data for reconstructed keypoint derived from keypoint network output.

This class is made by the KeypointReco class.

Public Functions

KPCluster()
virtual ~KPCluster()
void printInfo() const

Print cluster info to standard out.

Public Members

std::vector<float> center_pt_v

center point

std::vector<std::vector<float>> pt_pos_v

points associated to the center

std::vector<float> pt_score_v

score of cluster points

std::vector<std::vector<float>> pca_axis_v

principle component axes made by cluster_pca()

std::vector<float> pca_center

center of cluster

std::vector<float> pca_eigenvalues

principle component eigenvalues

std::vector<std::vector<float>> pca_ends_v

points on 1st pca-line out to the maximum projection distance from center

std::vector<std::vector<float>> bbox_v

axis-aligned bounding box. calculated along with pca

float pca_max_r

max distance from first principle component

float pca_ave_r2

average distance-squared from the first principle comonent

float pca_len

length of first principle component

int max_idx

hit in cluster with maximum score

float max_score

maximum score of hit in cluster

std::vector<float> max_pt_v

position of maximum point

int _cluster_idx

associated cluster_t in KeypointReco::_cluster_v (internal use only)

int _cluster_type

type of keypoint