Point Cloud Library (PCL)  1.9.1-dev
List of all members | Public Member Functions | Protected Member Functions | Protected Attributes
pcl::SVMClassify Class Reference

SVM (Support Vector Machines) classification of a dataset. More...

#include <pcl/ml/svm_wrapper.h>

+ Inheritance diagram for pcl::SVMClassify:

Public Member Functions

 SVMClassify ()
 Constructor. More...
 
 ~SVMClassify ()
 Destructor. More...
 
void setInputTrainingSet (std::vector< SVMData > training_set)
 It adds/store the training set with labelled data. More...
 
std::vector< SVMDatagetInputTrainingSet ()
 Return the current training set. More...
 
void resetTrainingSet ()
 Reset the training set. More...
 
bool loadClassifierModel (const char *filename)
 Read in a classifier model (in svmlight format). More...
 
void getClassificationResult (std::vector< std::vector< double > > &out)
 Get the result of the classification. More...
 
void saveClassificationResult (const char *filename)
 Save the classification result in an extern file. More...
 
void setClassifierModel (SVMModel model)
 Set the classifier model. More...
 
bool loadClassProblem (const char *filename)
 Read in a raw classification problem (in svmlight format). More...
 
bool loadNormClassProblem (const char *filename)
 Read in a normalized classification problem (in svmlight format). More...
 
void setProbabilityEstimates (bool set)
 Set whether the classification has to be done with the probability estimate. More...
 
bool classificationTest ()
 Start the classification on labelled input dataset. More...
 
bool classification ()
 Start the classification on un-labelled input dataset. More...
 
std::vector< double > classification (SVMData in)
 Start the classification on a single set. More...
 
bool saveClassProblem (const char *filename)
 Save the raw classification problem in a file (in svmlight format). More...
 
bool saveNormClassProblem (const char *filename)
 Save the normalized classification problem in a file (in svmlight format). More...
 
- Public Member Functions inherited from pcl::SVM
 SVM ()
 Constructor. More...
 
 ~SVM ()
 Destructor. More...
 
void getLabel (std::vector< int > &labels)
 Return the labels order from the classifier model. More...
 
void saveClassifierModel (const char *filename)
 Save the classifier model in an extern file (in svmlight format). More...
 

Protected Member Functions

void scaleProblem (svm_problem &input, svm_scaling scaling)
 It scales the input dataset using the model information. More...
 
- Protected Member Functions inherited from pcl::SVM
char * readline (FILE *input)
 To read a line from the input file. More...
 
void exitInputError (int line_num)
 Outputs an error in file reading. More...
 
const std::string & getClassName () const
 Get a string representation of the name of this class. More...
 
void adaptInputToLibSVM (std::vector< SVMData > training_set, svm_problem &prob)
 Convert the input format (vector of SVMData) into a readable format for libSVM. More...
 
void adaptLibSVMToInput (std::vector< SVMData > &training_set, svm_problem prob)
 Convert the libSVM format (svm_problem) into a easier output format. More...
 
bool loadProblem (const char *filename, svm_problem &prob)
 Load a problem from an extern file. More...
 
bool saveProblem (const char *filename, bool labelled)
 Save the raw problem in an extern file. More...
 
bool saveProblemNorm (const char *filename, svm_problem prob_, bool labelled)
 Save the problem (with normalized values) in an extern file. More...
 

Protected Attributes

bool model_extern_copied_
 
bool predict_probability_
 
std::vector< std::vector< double > > prediction_
 
- Protected Attributes inherited from pcl::SVM
std::vector< SVMDatatraining_set_
 
svm_problem prob_
 
SVMModel model_
 
svm_scaling scaling_
 
SVMParam param_
 
std::string class_name_
 
char * line_
 
int max_line_len_
 
bool labelled_training_set_
 

Additional Inherited Members

- Static Protected Member Functions inherited from pcl::SVM
static void printNull (const char *)
 Set for output printings during classification. More...
 

Detailed Description

SVM (Support Vector Machines) classification of a dataset.

It can be used both for testing a classifier model and for classify of new data.

Definition at line 360 of file svm_wrapper.h.

Constructor & Destructor Documentation

pcl::SVMClassify::SVMClassify ( )
inline

Constructor.

Definition at line 382 of file svm_wrapper.h.

pcl::SVMClassify::~SVMClassify ( )
inline

Destructor.

Definition at line 388 of file svm_wrapper.h.

Member Function Documentation

bool pcl::SVMClassify::classification ( )

Start the classification on un-labelled input dataset.

To get the classification result, use getClassificationResult.

Returns
false if fails.
std::vector<double> pcl::SVMClassify::classification ( SVMData  in)

Start the classification on a single set.

bool pcl::SVMClassify::classificationTest ( )

Start the classification on labelled input dataset.

It returns the accuracy percentage. To get the classification result, use getClassificationResult.

Returns
false if fails.
void pcl::SVMClassify::getClassificationResult ( std::vector< std::vector< double > > &  out)
inline

Get the result of the classification.

Definition at line 432 of file svm_wrapper.h.

std::vector<SVMData> pcl::SVMClassify::getInputTrainingSet ( )
inline

Return the current training set.

Definition at line 413 of file svm_wrapper.h.

bool pcl::SVMClassify::loadClassifierModel ( const char *  filename)

Read in a classifier model (in svmlight format).

Returns
false if fails.
bool pcl::SVMClassify::loadClassProblem ( const char *  filename)
inline

Read in a raw classification problem (in svmlight format).

The values are normalized using the classifier model information.

Returns
false if fails.

Definition at line 470 of file svm_wrapper.h.

References pcl::SVM::adaptLibSVMToInput(), and pcl::SVM::loadProblem().

bool pcl::SVMClassify::loadNormClassProblem ( const char *  filename)
inline

Read in a normalized classification problem (in svmlight format).

The data are kept whitout normalizing.

Returns
false if fails.

Definition at line 484 of file svm_wrapper.h.

References pcl::SVM::adaptLibSVMToInput(), and pcl::SVM::loadProblem().

void pcl::SVMClassify::resetTrainingSet ( )
inline

Reset the training set.

Definition at line 420 of file svm_wrapper.h.

void pcl::SVMClassify::saveClassificationResult ( const char *  filename)

Save the classification result in an extern file.

bool pcl::SVMClassify::saveClassProblem ( const char *  filename)
inline

Save the raw classification problem in a file (in svmlight format).

Returns
false if fails.

Definition at line 518 of file svm_wrapper.h.

References pcl::SVM::saveProblem().

bool pcl::SVMClassify::saveNormClassProblem ( const char *  filename)
inline

Save the normalized classification problem in a file (in svmlight format).

Returns
false if fails.

Definition at line 526 of file svm_wrapper.h.

References pcl::SVM::saveProblemNorm().

void pcl::SVMClassify::scaleProblem ( svm_problem input,
svm_scaling  scaling 
)
protected

It scales the input dataset using the model information.

void pcl::SVMClassify::setClassifierModel ( SVMModel  model)
inline

Set the classifier model.

Definition at line 444 of file svm_wrapper.h.

void pcl::SVMClassify::setInputTrainingSet ( std::vector< SVMData training_set)
inline

It adds/store the training set with labelled data.

Definition at line 396 of file svm_wrapper.h.

References pcl::SVM::adaptInputToLibSVM().

void pcl::SVMClassify::setProbabilityEstimates ( bool  set)
inline

Set whether the classification has to be done with the probability estimate.

(the classifier model has to support it).

Definition at line 494 of file svm_wrapper.h.

Member Data Documentation

bool pcl::SVMClassify::model_extern_copied_
protected

Definition at line 373 of file svm_wrapper.h.

bool pcl::SVMClassify::predict_probability_
protected

Definition at line 374 of file svm_wrapper.h.

std::vector< std::vector<double> > pcl::SVMClassify::prediction_
protected

Definition at line 375 of file svm_wrapper.h.


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