mrpt::vision::CFeatureExtraction Class Reference

The central class from which images can be analyzed in search of different kinds of interest points and descriptors computed for them. More...

#include <mrpt/vision/CFeatureExtraction.h>

List of all members.

Classes
struct	TOptions
	The set of parameters for all the detectors & descriptor algorithms. More...
Public Types
enum	TSIFTImplementation { LoweBinary = 0, CSBinary, VedaldiBinary, Hess }
Public Member Functions
	CFeatureExtraction ()
	Constructor.
virtual	~CFeatureExtraction ()
	Virtual destructor.
void	detectFeatures (const CImage &img, CFeatureList &feats, unsigned int init_ID=0, unsigned int nDesiredFeatures=0, const TImageROI &ROI=TImageROI()) const
	Extract features from the image based on the method defined in TOptions.
void	computeDescriptors (const CImage &in_img, CFeatureList &inout_features, TDescriptorType in_descriptor_list) const
	Compute one (or more) descriptors for the given set of interest points onto the image, which may have been filled out manually or from detectFeatures.
void	findMoreFeatures (const CImage &img, const CFeatureList &inList, CFeatureList &outList, unsigned int nDesiredFeats=0) const
	Extract more features from the image (apart from the provided ones) based on the method defined in TOptions.
Public Attributes
TOptions	options
	Set all the parameters of the desired method here before calling "detectFeatures".
Private Member Functions
void	internal_computeSiftDescriptors (const CImage &in_img, CFeatureList &in_features) const
	Compute the SIFT descriptor of the provided features into the input image.
void	internal_computeSurfDescriptors (const CImage &in_img, CFeatureList &in_features) const
	Compute the SURF descriptor of the provided features into the input image.
void	internal_computeSpinImageDescriptors (const CImage &in_img, CFeatureList &in_features) const
	Compute the intensity-domain spin images descriptor of the provided features into the input image.
void	internal_computePolarImageDescriptors (const CImage &in_img, CFeatureList &in_features) const
	Compute a polar-image descriptor of the provided features into the input image.
void	internal_computeLogPolarImageDescriptors (const CImage &in_img, CFeatureList &in_features) const
	Compute a log-polar image descriptor of the provided features into the input image.
void	selectGoodFeaturesKLT (const CImage &inImg, CFeatureList &feats, unsigned int init_ID=0, unsigned int nDesiredFeatures=0, void *mask_=NULL) const
	Select good features using the openCV implementation of the KLT method.
void	extractFeaturesKLT (const CImage &img, CFeatureList &feats, unsigned int init_ID=0, unsigned int nDesiredFeatures=0, const TImageROI &ROI=TImageROI()) const
	Extract features from the image based on the KLT method.
void	extractFeaturesBCD (const CImage &img, CFeatureList &feats, unsigned int init_ID=0, unsigned int nDesiredFeatures=0, const TImageROI &ROI=TImageROI()) const
	Extract features from the image based on the BCD method.
void	extractFeaturesSIFT (const CImage &img, CFeatureList &feats, unsigned int init_ID=0, unsigned int nDesiredFeatures=0, const TImageROI &ROI=TImageROI()) const
	Extract features from the image based on the SIFT method.
void	extractFeaturesSURF (const CImage &img, CFeatureList &feats, unsigned int init_ID=0, unsigned int nDesiredFeatures=0, const TImageROI &ROI=TImageROI()) const
	Extract features from the image based on the SURF method.
void *	my_scale_space_extrema (CFeatureList &featList, void *dog_pyr, int octvs, int intvls, double contr_thr, int curv_thr, void *storage) const
	Computes extrema in the scale space.
void	my_adjust_for_img_dbl (void *features) const
	Adjust scale if the image was initially doubled.
void	getTimesExtrema (void *dog_pyr, int octvs, int intvls, float row, float col, unsigned int &nMin, unsigned int &nMax) const
	Gets the number of times that a point in the image is higher or lower than the surroundings in the image-scale space.
double	getLaplacianValue (void *dog_pyr, int octvs, int intvls, float row, float col) const
	Computes the Laplacian value of the feature in the corresponing image in the pyramid.
void	insertCvSeqInCFeatureList (void *features, CFeatureList &list, unsigned int init_ID=0) const
	Append a sequence of openCV features into an MRPT feature list.
void	convertCvSeqInCFeatureList (void *features, CFeatureList &list, unsigned int init_ID=0, const TImageROI &ROI=TImageROI()) const
	Converts a sequence of openCV features into an MRPT feature list.

---

Detailed Description

The central class from which images can be analyzed in search of different kinds of interest points and descriptors computed for them.

To extract features from an image, create an instance of CFeatureExtraction, fill out its CFeatureExtraction::options field, including the algorithm to use (see CFeatureExtraction::TMethodExtraction), and call CFeatureExtraction::detectFeatures. This will return a set of features of the class mrpt::vision::CFeature, which include details for each interest point as well as the desired descriptors and/or patches.

By default, a 21x21 patch is extracted for each detected feature. If the patch is not needed, set patchSize to 0 in CFeatureExtraction::options

The implemented detection algorithms are (see CFeatureExtraction::TMethodExtraction):

• KLT (Kanade-Lucas-Tomasi): A detector (no descriptor vector).
• Harris: A detector (no descriptor vector).
• BCD (Binary Corner Detector): A detector (no descriptor vector) (Not implemented yet).
• SIFT: An implementation of the SIFT detector and descriptor. The implemention may be selected with CFeatureExtraction::TOptions::SIFTOptions::implementation.
• SURF: OpenCV's implementation of SURF detector and descriptor.

Additionally, given a list of interest points onto an image, the following descriptors can be computed for each point by calling CFeatureExtraction::computeDescriptors :

• SIFT descriptor (Lowe's descriptors).
• SURF descriptor (OpenCV's implementation - Requires OpenCV 1.1.0 from SVN or later).
• Intensity-domain spin images (SpinImage): Creates a vector descriptor with the 2D histogram as a single row.
• A circular patch in polar coordinates (Polar images): The matrix descriptor is a 2D polar image centered at the interest point.
• A log-polar image patch (Log-polar images): The matrix descriptor is the 2D log-polar image centered at the interest point.

Note:

The descriptor "Intensity-domain spin images" is described in "A sparse texture representation using affine-invariant regions", S Lazebnik, C Schmid, J Ponce, 2003 IEEE Computer Society Conference on Computer Vision.

See also:

mrpt::vision::CFeature

Definition at line 70 of file CFeatureExtraction.h.

---

Member Enumeration Documentation

enum mrpt::vision::CFeatureExtraction::TSIFTImplementation

Enumerator:

LoweBinary
CSBinary
VedaldiBinary
Hess

Definition at line 73 of file CFeatureExtraction.h.

---

Constructor & Destructor Documentation

mrpt::vision::CFeatureExtraction::CFeatureExtraction

(

)

Constructor.

virtual mrpt::vision::CFeatureExtraction::~CFeatureExtraction

(

)

[virtual]

Virtual destructor.

---

Member Function Documentation

void mrpt::vision::CFeatureExtraction::computeDescriptors	(	const CImage &	in_img,
		CFeatureList &	inout_features,
		TDescriptorType	in_descriptor_list
	)			const

Compute one (or more) descriptors for the given set of interest points onto the image, which may have been filled out manually or from detectFeatures.

Parameters:

	in_img	(input) The image from where to compute the descriptors.
	inout_features	(input/output) The list of features whose descriptors are going to be computed.
	in_descriptor_list	(input) The bitwise OR of one or several descriptors defined in TDescriptorType.

Each value in "in_descriptor_list" represents one descriptor to be computed, for example:

    // This call will compute both, SIFT and Spin-Image descriptors for a list of feature points lstFeats.
    fext.computeDescriptors(img, lstFeats, descSIFT | descSpinImages );

Note:

The SIFT descriptors for already located features can only be computed through the Hess and CSBinary implementations which may be specified in CFeatureExtraction::TOptions::SIFTOptions.

This call will also use additional parameters from options

void mrpt::vision::CFeatureExtraction::convertCvSeqInCFeatureList	(	void *	features,
		CFeatureList &	list,
		unsigned int	init_ID = 0,
		const TImageROI &	ROI = TImageROI()
	)			const [private]

Converts a sequence of openCV features into an MRPT feature list.

Parameters:

	features	The sequence of features.
	list	[in-out] The list of MRPT features.
	init_ID	[in][optional] The initial ID for the features (default = 0).
	ROI	[in][optional] The initial ID for the features (default = empty ROI -> not used).

void mrpt::vision::CFeatureExtraction::detectFeatures	(	const CImage &	img,
		CFeatureList &	feats,
		unsigned int	init_ID = 0,
		unsigned int	nDesiredFeatures = 0,
		const TImageROI &	ROI = TImageROI()
	)			const

Extract features from the image based on the method defined in TOptions.

Parameters:

	img	(input) The image from where to extract the images.
	feats	(output) A complete list of features (containing a patch for each one of them if options.patchsize > 0).
	nDesiredFeatures	(op. input) Number of features to be extracted. Default: all possible.
	ROI	(op. input) Region of Interest. Default: The whole image.

See also:

computeDescriptors

void mrpt::vision::CFeatureExtraction::extractFeaturesBCD	(	const CImage &	img,
		CFeatureList &	feats,
		unsigned int	init_ID = 0,
		unsigned int	nDesiredFeatures = 0,
		const TImageROI &	ROI = TImageROI()
	)			const [private]

Extract features from the image based on the BCD method.

Parameters:

	img	The image from where to extract the images.
	feats	The list of extracted features.
	nDesiredFeatures	Number of features to be extracted. Default: authomatic.
	ROI	(op. input) Region of Interest. Default: All the image.

void mrpt::vision::CFeatureExtraction::extractFeaturesKLT	(	const CImage &	img,
		CFeatureList &	feats,
		unsigned int	init_ID = 0,
		unsigned int	nDesiredFeatures = 0,
		const TImageROI &	ROI = TImageROI()
	)			const [private]

Extract features from the image based on the KLT method.

Parameters:

	img	The image from where to extract the images.
	feats	The list of extracted features.
	nDesiredFeatures	Number of features to be extracted. Default: authomatic.
	ROI	(op. input) Region of Interest. Default: All the image.

void mrpt::vision::CFeatureExtraction::extractFeaturesSIFT	(	const CImage &	img,
		CFeatureList &	feats,
		unsigned int	init_ID = 0,
		unsigned int	nDesiredFeatures = 0,
		const TImageROI &	ROI = TImageROI()
	)			const [private]

Extract features from the image based on the SIFT method.

Parameters:

	img	The image from where to extract the images.
	feats	The list of extracted features.
	nDesiredFeatures	Number of features to be extracted. Default: authomatic.
	ROI	(op. input) Region of Interest. Default: All the image.

void mrpt::vision::CFeatureExtraction::extractFeaturesSURF	(	const CImage &	img,
		CFeatureList &	feats,
		unsigned int	init_ID = 0,
		unsigned int	nDesiredFeatures = 0,
		const TImageROI &	ROI = TImageROI()
	)			const [private]

Extract features from the image based on the SURF method.

Parameters:

	img	The image from where to extract the images.
	feats	The list of extracted features.
	nDesiredFeatures	Number of features to be extracted. Default: authomatic.
	ROI	(op. input) Region of Interest. Default: All the image.

void mrpt::vision::CFeatureExtraction::findMoreFeatures	(	const CImage &	img,
		const CFeatureList &	inList,
		CFeatureList &	outList,
		unsigned int	nDesiredFeats = 0
	)			const

Extract more features from the image (apart from the provided ones) based on the method defined in TOptions.

Parameters:

	img	(input) The image from where to extract the images.
	inList	(input) The actual features in the image.
	outList	(output) The list of new features (containing a patch for each one of them if options.patchsize > 0).
	nDesiredFeatures	(op. input) Number of features to be extracted. Default: all possible.

double mrpt::vision::CFeatureExtraction::getLaplacianValue	(	void *	dog_pyr,
		int	octvs,
		int	intvls,
		float	row,
		float	col
	)			const [private]

Computes the Laplacian value of the feature in the corresponing image in the pyramid.

Parameters:

	dog_pyr	Pyramid of images.
	octvs	Number of considered octaves.
	intvls	Number of intervales in octaves.
	row	The row of the feature in the original image.
	col	The column of the feature in the original image.

void mrpt::vision::CFeatureExtraction::getTimesExtrema	(	void *	dog_pyr,
		int	octvs,
		int	intvls,
		float	row,
		float	col,
		unsigned int &	nMin,
		unsigned int &	nMax
	)			const [private]

Gets the number of times that a point in the image is higher or lower than the surroundings in the image-scale space.

Parameters:

	dog_pyr	Pyramid of images.
	octvs	Number of considered octaves.
	intvls	Number of intervales in octaves.
	row	The row of the feature in the original image.
	col	The column of the feature in the original image.
	nMin	[out]: Times that the feature is lower than the surroundings.
	nMax	[out]: Times that the feature is higher than the surroundings.

void mrpt::vision::CFeatureExtraction::insertCvSeqInCFeatureList	(	void *	features,
		CFeatureList &	list,
		unsigned int	init_ID = 0
	)			const [private]

Append a sequence of openCV features into an MRPT feature list.

Parameters:

	features	The sequence of features.
	list	[in-out] The list of MRPT features.
	init_ID	[in] The initial ID for the new features.

void mrpt::vision::CFeatureExtraction::internal_computeLogPolarImageDescriptors	(	const CImage &	in_img,
		CFeatureList &	in_features
	)			const [private]

Compute a log-polar image descriptor of the provided features into the input image.

Parameters:

	in_img	(input) The image from where to compute the descriptors.
	in_features	(input/output) The list of features whose descriptors are going to be computed.

Note:

Additional parameters from CFeatureExtraction::TOptions::LogPolarImagesOptions are used in this method.

void mrpt::vision::CFeatureExtraction::internal_computePolarImageDescriptors	(	const CImage &	in_img,
		CFeatureList &	in_features
	)			const [private]

Compute a polar-image descriptor of the provided features into the input image.

Parameters:

	in_img	(input) The image from where to compute the descriptors.
	in_features	(input/output) The list of features whose descriptors are going to be computed.

Note:

Additional parameters from CFeatureExtraction::TOptions::PolarImagesOptions are used in this method.

void mrpt::vision::CFeatureExtraction::internal_computeSiftDescriptors	(	const CImage &	in_img,
		CFeatureList &	in_features
	)			const [private]

Compute the SIFT descriptor of the provided features into the input image.

Parameters:

	in_img	(input) The image from where to compute the descriptors.
	in_features	(input/output) The list of features whose descriptors are going to be computed.

Note:

The SIFT descriptors for already located features can only be computed through the Hess and CSBinary implementations which may be specified in CFeatureExtraction::TOptions::SIFTOptions.

void mrpt::vision::CFeatureExtraction::internal_computeSpinImageDescriptors	(	const CImage &	in_img,
		CFeatureList &	in_features
	)			const [private]

Compute the intensity-domain spin images descriptor of the provided features into the input image.

Parameters:

	in_img	(input) The image from where to compute the descriptors.
	in_features	(input/output) The list of features whose descriptors are going to be computed.

Note:

Additional parameters from CFeatureExtraction::TOptions::SpinImagesOptions are used in this method.

void mrpt::vision::CFeatureExtraction::internal_computeSurfDescriptors	(	const CImage &	in_img,
		CFeatureList &	in_features
	)			const [private]

Compute the SURF descriptor of the provided features into the input image.

Parameters:

	in_img	(input) The image from where to compute the descriptors.
	in_features	(input/output) The list of features whose descriptors are going to be computed.

void mrpt::vision::CFeatureExtraction::my_adjust_for_img_dbl

(

void *

features

)

const [private]

Adjust scale if the image was initially doubled.

Parameters:

features

The sequence of features.

void* mrpt::vision::CFeatureExtraction::my_scale_space_extrema	(	CFeatureList &	featList,
		void *	dog_pyr,
		int	octvs,
		int	intvls,
		double	contr_thr,
		int	curv_thr,
		void *	storage
	)			const [private]

Computes extrema in the scale space.

Parameters:

	dog_pyr	Pyramid of images.
	octvs	Number of considered octaves.
	intvls	Number of intervales in octaves.

void mrpt::vision::CFeatureExtraction::selectGoodFeaturesKLT	(	const CImage &	inImg,
		CFeatureList &	feats,
		unsigned int	init_ID = 0,
		unsigned int	nDesiredFeatures = 0,
		void *	mask_ = NULL
	)			const [private]

Select good features using the openCV implementation of the KLT method.

Parameters:

	img	(input) The image from where to select extract the images.
	feats	(output) A complete list of features (containing a patch for each one of them if options.patchsize > 0).
	nDesiredFeatures	(op. input) Number of features to be extracted. Default: all possible.
	omitPixels	(op. input) A mask for determining the ROI. (0: do not omit this pixel, 1: omit this pixel)

---

Member Data Documentation

TOptions mrpt::vision::CFeatureExtraction::options

Set all the parameters of the desired method here before calling "detectFeatures".

Definition at line 181 of file CFeatureExtraction.h.