Point Cloud Library (PCL)  1.14.1-dev
List of all members | Public Member Functions | Protected Attributes
pcl::people::HOG Class Reference

HOG represents a class for computing the HOG descriptor described in Dalal, N. More...

#include <pcl/people/hog.h>

Public Member Functions

 HOG ()
 Constructor. More...
 
virtual ~HOG ()
 Destructor. More...
 
void gradMag (float *I, int h, int w, int d, float *M, float *O) const
 Compute gradient magnitude and orientation at each location (uses sse). More...
 
void gradHist (float *M, float *O, int h, int w, int bin_size, int n_orients, bool soft_bin, float *H) const
 Compute n_orients gradient histograms per bin_size x bin_size block of pixels. More...
 
void normalization (float *H, int h, int w, int bin_size, int n_orients, float clip, float *G) const
 Normalize histogram of gradients. More...
 
void compute (float *I, int h, int w, int n_channels, int bin_size, int n_orients, bool soft_bin, float *descriptor)
 Compute HOG descriptor. More...
 
void compute (float *I, float *descriptor) const
 Compute HOG descriptor with default parameters. More...
 

Protected Attributes

int h_
 image height (default = 128) More...
 
int w_
 image width (default = 64) More...
 
int n_channels_
 image number of channels (default = 3) More...
 
int bin_size_
 spatial bin size (default = 8) More...
 
int n_orients_
 number of orientation bins (default = 9) More...
 
bool soft_bin_
 if true, each pixel can contribute to multiple spatial bins (using bilinear interpolation) (default = true) More...
 
float clip_
 value at which to clip histogram bins (default = 0.2) More...
 

Detailed Description

HOG represents a class for computing the HOG descriptor described in Dalal, N.

and Triggs, B., "Histograms of oriented gradients for human detection", CVPR 2005.

Author
Matteo Munaro, Stefano Ghidoni, Stefano Michieletto

Definition at line 54 of file hog.h.

Constructor & Destructor Documentation

◆ HOG()

pcl::people::HOG::HOG ( )

Constructor.

◆ ~HOG()

virtual pcl::people::HOG::~HOG ( )
virtual

Destructor.

Member Function Documentation

◆ compute() [1/2]

void pcl::people::HOG::compute ( float *  I,
float *  descriptor 
) const

Compute HOG descriptor with default parameters.

Parameters
[in]IImage as array of float between 0 and 1.
[out]descriptorHOG descriptor.

◆ compute() [2/2]

void pcl::people::HOG::compute ( float *  I,
int  h,
int  w,
int  n_channels,
int  bin_size,
int  n_orients,
bool  soft_bin,
float *  descriptor 
)

Compute HOG descriptor.

Parameters
[in]IImage as array of float between 0 and 1.
[in]hImage height.
[in]wImage width.
[in]n_channelsImage number of channels.
[in]bin_sizeSpatial bin size.
[in]n_orientsNumber of orientation bins.
[in]soft_binIf true, each pixel can contribute to multiple spatial bins (using bilinear interpolation).
[out]descriptorHOG descriptor.

Referenced by pcl::people::PersonClassifier< PointT >::evaluate().

◆ gradHist()

void pcl::people::HOG::gradHist ( float *  M,
float *  O,
int  h,
int  w,
int  bin_size,
int  n_orients,
bool  soft_bin,
float *  H 
) const

Compute n_orients gradient histograms per bin_size x bin_size block of pixels.


Parameters
[in]MGradient magnitude for each image point.
[in]OGradient orientation for each image point.
[in]hImage height.
[in]wImage width.
[in]bin_sizeSpatial bin size.
[in]n_orientsNumber of orientation bins.
[in]soft_binIf true, each pixel can contribute to multiple spatial bins (using bilinear interpolation).
[out]HGradient histograms.

◆ gradMag()

void pcl::people::HOG::gradMag ( float *  I,
int  h,
int  w,
int  d,
float *  M,
float *  O 
) const

Compute gradient magnitude and orientation at each location (uses sse).

Parameters
[in]IImage as array of float.
[in]hImage height.
[in]wImage width.
[in]dImage number of channels.
[out]MGradient magnitude for each image point.
[out]OGradient orientation for each image point.

◆ normalization()

void pcl::people::HOG::normalization ( float *  H,
int  h,
int  w,
int  bin_size,
int  n_orients,
float  clip,
float *  G 
) const

Normalize histogram of gradients.

Parameters
[in]HGradient histograms.
[in]hImage height.
[in]wImage width.
[in]bin_sizeSpatial bin size.
[in]n_orientsNumber of orientation bins.
[in]clipValue at which to clip histogram bins.
[out]GNormalized gradient histograms.

Member Data Documentation

◆ bin_size_

int pcl::people::HOG::bin_size_
protected

spatial bin size (default = 8)

Definition at line 174 of file hog.h.

◆ clip_

float pcl::people::HOG::clip_
protected

value at which to clip histogram bins (default = 0.2)

Definition at line 183 of file hog.h.

◆ h_

int pcl::people::HOG::h_
protected

image height (default = 128)

Definition at line 165 of file hog.h.

◆ n_channels_

int pcl::people::HOG::n_channels_
protected

image number of channels (default = 3)

Definition at line 171 of file hog.h.

◆ n_orients_

int pcl::people::HOG::n_orients_
protected

number of orientation bins (default = 9)

Definition at line 177 of file hog.h.

◆ soft_bin_

bool pcl::people::HOG::soft_bin_
protected

if true, each pixel can contribute to multiple spatial bins (using bilinear interpolation) (default = true)

Definition at line 180 of file hog.h.

◆ w_

int pcl::people::HOG::w_
protected

image width (default = 64)

Definition at line 168 of file hog.h.


The documentation for this class was generated from the following file: