#include <Filter.h>

Inheritance diagram for GRT::Filter:

Public Member Functions
	Filter (const char *classType, UINT filterSize=5, UINT numDimensions=1)

	Filter (const Filter &rhs)

virtual	~Filter ()

Filter &	operator= (const Filter &rhs)

virtual bool	deepCopyFrom (const PreProcessing *preProcessing)

virtual bool	process (const VectorDouble &inputVector)

virtual bool	reset ()

virtual bool	saveModelToFile (string filename) const

virtual bool	saveModelToFile (fstream &file) const

virtual bool	loadModelFromFile (string filename)

virtual bool	loadModelFromFile (fstream &file)

bool	init (UINT filterSize, UINT numDimensions)

double	filter (const double x)

VectorDouble	filter (const VectorDouble &x)

virtual double	computeFilter (const VectorDouble &buf)=0

UINT	getFilterSize () const

VectorDouble	getFilteredData () const

vector< VectorDouble >	getDataBuffer () const

Protected Attributes
UINT	filterSize
	The size of the filter. More...

UINT	inputSampleCounter
	A counter to keep track of the number of input samples. More...

CircularBuffer< VectorDouble >	dataBuffer
	A buffer to store the previous N values, N = filterSize. More...

Constructor & Destructor Documentation

GRT::Filter::Filter	(	const char *	classType,
		UINT	filterSize = `5`,
		UINT	numDimensions = `1`
	)

Constructor, sets the size of the median filter and the dimensionality of the data it will filter.

Parameters

char	*classType: the name of the sub-class being instantiated.
UINT	filterSize: the size of the median filter, should be a value greater than zero. Default filterSize = 5
UINT	numDimensions: the dimensionality of the data to filter. Default numDimensions = 1

GRT::Filter::Filter ( const Filter & rhs )

Copy Constructor, copies the Filter from the rhs instance to this instance

Parameters

const Filter &rhs: another instance of the Filter class from which the data will be copied to this instance

GRT::Filter::~Filter ( )

virtual

Default Destructor

Member Function Documentation

virtual double GRT::Filter::computeFilter ( const VectorDouble & buf )

pure virtual

Compute the filter. This is implemented by the derived class and called separately on each dimension of the input data.

Parameters

const VectorDouble buf: the values to filter, i.e. the last filterSize (or fewer) values read within a particular dimension

Returns: the filtered value

Implemented in LogEnergy.

bool GRT::Filter::deepCopyFrom ( const PreProcessing * preProcessing )

virtual

Sets the PreProcessing deepCopyFrom function, overwriting the base PreProcessing function. This function is used to deep copy the values from the input pointer to this instance of the PreProcessing module. This function is called by the GestureRecognitionPipeline when the user adds a new PreProcessing module to the pipeline.

Parameters

const PreProcessing *preProcessing: a pointer to another instance of a Filter, the values of that instance will be cloned to this instance

Returns: true if the deep copy was successful, false otherwise

double GRT::Filter::filter ( const double x )

Filters the input, this should only be called if the dimensionality of the filter was set to 1.

Parameters

double x: the value to filter, this should only be called if the dimensionality of the filter was set to 1

Returns: the filtered value. Zero will be returned if the value was not filtered

VectorDouble GRT::Filter::filter ( const VectorDouble & x )

Filters the input, the dimensionality of the input vector should match that of the filter.

Parameters

const VectorDouble &x: the values to filter, the dimensionality of the input vector should match that of the filter

Returns: the filtered values. An empty vector will be returned if the values were not filtered

vector< VectorDouble > GRT::Filter::getDataBuffer ( ) const

Returns the current data in the dataBuffer. This will be a [N M] vector of VectorDoubles, where N is the number of dimensions in the filter and M is the size of the buffer. If the filter has not been initialized then an empty vector will be returned.

Returns: the current data in the dataBuffer

VectorDouble GRT::Filter::getFilteredData ( ) const

inline

Returns the last value(s) that were filtered.

Returns: the filtered values. An empty vector will be returned if the values were not filtered

UINT GRT::Filter::getFilterSize ( ) const

inline

Gets the current filter size.

Returns: returns the filter size

bool GRT::Filter::init	(	UINT	filterSize,
		UINT	numDimensions
	)

Initializes the filter, setting the filter size and dimensionality of the data it will filter. Sets all the filter values to zero.

Parameters

UINT	filterSize: the size of the moving average filter, should be a value greater than zero

Returns: true if the filter was initiliazed, false otherwise

bool GRT::Filter::loadModelFromFile ( string filename )

virtual

This loads the Filter settings from a file. This overrides the loadModelFromFile function in the PreProcessing base class.

Parameters

string filename: the name of the file to load the settings from

Returns: returns true if the settings were loaded successfully, false otherwise

bool GRT::Filter::loadModelFromFile ( fstream & file )

virtual

This loads the Filter settings from a file. This overrides the loadModelFromFile function in the PreProcessing base class.

Parameters

fstream &file: a reference to the file to load the settings from

Returns: returns true if the model was loaded successfully, false otherwise

Filter & GRT::Filter::operator= ( const Filter & rhs )

Sets the equals operator, copies the data from the rhs instance to this instance

Parameters

const Filter &rhs: another instance of the Filter class from which the data will be copied to this instance

Returns: a reference to this instance of Filter

bool GRT::Filter::process ( const VectorDouble & inputVector )

virtual

Sets the PreProcessing process function, overwriting the base PreProcessing function. This function is called by the GestureRecognitionPipeline when any new input data needs to be processed (during the prediction phase for example). This function calls the Filter's filter function.

Parameters

const VectorDouble &inputVector: the inputVector that should be processed. Must have the same dimensionality as the PreProcessing module

Returns: true if the data was processed, false otherwise

bool GRT::Filter::reset ( )

virtual

Sets the PreProcessing reset function, overwriting the base PreProcessing function. This function is called by the GestureRecognitionPipeline when the pipelines main reset() function is called. This function resets the filter values by re-initiliazing the filter.

Returns: true if the filter was reset, false otherwise

bool GRT::Filter::saveModelToFile ( string filename ) const

virtual

This saves the current settings of the Filter to a file. This overrides the saveModelToFile function in the PreProcessing base class.

Parameters

string filename: the name of the file to save the settings to

Returns: returns true if the model was saved successfully, false otherwise

bool GRT::Filter::saveModelToFile ( fstream & file ) const

virtual

This saves the current settings of the Filter to a file. This overrides the saveModelToFile function in the PreProcessing base class.

Parameters

fstream &file: a reference to the file the settings will be saved to

Returns: returns true if the settings were saved successfully, false otherwise

Member Data Documentation

CircularBuffer< VectorDouble > GRT::Filter::dataBuffer

protected

A buffer to store the previous N values, N = filterSize.

UINT GRT::Filter::filterSize

protected

The size of the filter.

UINT GRT::Filter::inputSampleCounter

protected

A counter to keep track of the number of input samples.

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

src/Filter.h
src/Filter.cpp

Public Member Functions

Protected Attributes

Constructor & Destructor Documentation

Member Function Documentation

Member Data Documentation