Classes | |
class | Bitmap |
class | Clusteror |
Base class for objects that cluster feature vectors. More... | |
class | ClusterorFactory |
class | Color64 |
class | Computor |
Base class for feature computors. More... | |
class | Concept |
class | ConceptSet |
class | FeatureDefinition |
class | FeatureParameter |
class | FeatureTable |
A table with features. More... | |
class | FeatureTableSet |
class | Grid |
class | HarrisLaplaceDetector |
class | InterestPointFeature |
class | IntWeibullNgbPnLoop |
Neighbourhood functor for weibull fitting. More... | |
class | FISTDescriptor |
class | LabelSet |
class | PointDescriptorTable |
A table with points and their descriptors. More... | |
class | RadiusClusteror |
Radius based clustering of feature vectors using similarity. More... | |
class | RandomTree |
Representation of a random tree (a decision tree that maps a feature vector to a code word). More... | |
class | TestFunctions |
This class tests some feature functionality. More... | |
class | TestMakeRandomTree |
class | TestRandomTree |
class | VisSem |
class | WeibullNgbPnLoop |
Neighbourhood functor for weibull fitting. More... | |
Typedefs | |
typedef std::vector< String > | FeatureList |
typedef Table::TableTem< Column::ColumnTem< Quid >, Vector::VectorSet< Array::Array2dScalarReal64 > > | FeatureTableBaseType |
typedef Table::TableTem< Vector::ColumnVectorSet, Column::ColumnInt32 > | AnnotatedFeatureTable |
typedef Table::TableTem< Column::ColumnTem< Real64 >, Column::ColumnTem< Real64 >, Column::ColumnTem< Real64 >, Column::ColumnTem< Real64 >, Column::ColumnTem< Real64 >, Vector::VectorSet< Array::Array2dScalarReal64 > > | PointDescriptorTableBaseType |
typedef std::vector< RandomTree * > | RandomForest |
typedef Table::TableTem< Column::ColumnInt32, Column::ColumnInt32, Column::ColumnReal64 > | RandomTreeTable |
Functions | |
ILOG_VAR_INIT (ClusterorFactory, Impala.Core.Feature) | |
ILOG_VAR_INIT (Computor, Impala.Core.Feature) | |
void | AddDSurfOptions (CmdOptions &options) |
void | GetDSurfOptions (CmdOptions &options, int &haarSize, int &subregionSize, int &spaciality) |
void | ExtractFeatureIntoVector (std::vector< Real64 > &vector, std::vector< Array::Array2dScalarReal64 * > &features, int x, int y, int spaciality) |
subroutine of DSurf - extract from the collection of images 'features'; at the point (x,y) a feature that we will store in vector 'v' x and y are assumed to be in range of the image considering spaciality | |
void | DSurf (Array::Array2dScalarReal64 *image, Geometry::InterestPointList &pointList, int sampleRate, int numberHaarSum, int spaciality) |
image in, descriptors are added to pointList | |
void | DumpFeatureTableHistogram (FeatureTable *table, FeatureDefinition fdef, int id) |
ILOG_VAR_INIT (FeatureTable, Impala.Core.Feature) | |
ILOG_VAR_INIT (FeatureTableSet, Impala.Core.Feature) | |
void | Normalize (Array::Array2dScalarReal64 *im) |
std::vector< Array::Array2dScalarReal64 * > | GetColorChannels (Array::Array2dVec3Real64 *input, String colorModel) |
std::vector< Array::Array2dScalarReal64 * > | GetColorChannels (Array::Array2dVec3UInt8 *inputInt, String colorModel) |
void | HaarFirstLine (std::vector< Array::Array2dScalarReal64 * > &response, Array::Array2dScalarReal64 *image, int y) |
void | HaarSecondLine (std::vector< Array::Array2dScalarReal64 * > &response, Array::Array2dScalarReal64 *image, int y) |
std::vector< Array::Array2dScalarReal64 * > | HaarResponse (Array::Array2dScalarReal64 *image) |
returns a vector of four images:
| |
ILOG_VAR_INIT (HarrisLaplaceDetector, Core.Feature) | |
ILOG_VAR_INIT (InterestPointFeature, Core.Feature) | |
bool | StringIsFullyNumeric (String s) |
void | CalculateFISTDescriptors (Array::Array2dVec3UInt8 *inputNoBorder, PointDescriptorTable *pointData, String descriptor) |
template<class SrcArrayT> | |
PointDescriptorTable * | LaplacianDetection (SrcArrayT *imIntensity, bool useRecGauss, Real64 precision) |
template<class SrcArrayT> | |
PointDescriptorTable * | LaplacianDetector (SrcArrayT *input, bool useRecGauss=true, Real64 precision=3.0) |
void | Dump (AnnotatedFeatureTable *t, std::ostream &os) |
Histogram::Histogram1dTem< int > * | MakeHistogram (const AnnotatedFeatureTable *data, int nrClasses, bool *filter) |
makes a historgram of the values found in the second column of data (the column with the 'class id') considering the filter. | |
double | Gain (const AnnotatedFeatureTable *data, int nrClasses, bool *left, bool *right) |
according to Jasper: The Gain is defined as in [Shotton08], taken again from Lepetit, CVPR 2005 | |
void | SplitSet (bool *&left, bool *&right, int dimension, double value, const AnnotatedFeatureTable *data, bool *filter) |
returns filters for left and right through ref to pointer (pointer is set to a new address, so if either of these point to data when this function is called it will introduce a memory leak). | |
void | TryRandomSplit (int &dimension, double &value, double &gain, const AnnotatedFeatureTable *data, bool *filter, int nrClasses) |
Take a random split point and return it and it's gain in the first three reference parameters. | |
void | FindSplit (int &dimension, double &value, const AnnotatedFeatureTable *data, bool *filter, int nrClasses, int nrTrials) |
Find a split by trying random splits and returning the one with the maximum gain. | |
int | GetCodeWord () |
returns a new id every time the function is called. | |
Feature::RandomTree * | MakeRandomTree (const AnnotatedFeatureTable *data, bool *filter, int nrClasses, int maxDepth, int nrTrials) |
recursive funtion finds a split and calls itself for the 'left' and 'right' splits | |
RandomTree * | MakeRandomTree (const AnnotatedFeatureTable *data, int nrClasses, int maxDepth, int nrTrials) |
void | Dump (Matrix::Mat *matrix) |
Vector::VectorTem< Real64 > | MarkovStationaryFeature (Matrix::Mat *cooccuranceMatrix) |
Vector::VectorTem< Real64 > | MarkovStationaryFeatureTimed (Matrix::Mat *m) |
void | testMSF () |
ILOG_VAR_INIT (PointDescriptorTable, Impala.Core.Feature) | |
RandomForest | ReadRandomForest (RandomTreeTable *table) |
RandomForest | ReadRandomForest (FeatureTable *table) |
int | GetCodebookLength (FeatureTable *table) |
void | DeleteForest (RandomForest forest) |
Or should we make this the method of a class RandomForest? | |
ILOG_VAR_INIT (RandomTree, Impala.Core.Feature) | |
void | Write (RandomTree *tree, RandomTreeTable *table) |
Writes the tree to the end of the table. | |
RandomTree * | Read (RandomTreeTable *table, int &index) |
Call with valid table and the index of the position in the tree where the tree starts (0 for example). | |
RandomTreeTable * | MakeRandomTreeTable (FeatureTable *ft) |
FeatureTable * | MakeFeatureTable (RandomTreeTable *rtt) |
Array::Array2dScalarReal64 * | CreateCircleMask (int patchSize) |
void | ComputeRegionDescriptor (std::vector< Real64 > &vec, Array::Array2dVec3UInt8 *src, String descriptor, bool useCircularMask, int srcCenterX, int srcCenterY) |
void | MakePatch (Array::Array2dVec3UInt8 *&patch, Array::Array2dVec3UInt8 *img, int x, int y, int pixelsAround) |
void | CalculateRegionDescriptors (Array::Array2dVec3UInt8 *inputNoBorder, Geometry::InterestPointList &pointList, String descriptor) |
void | CalculateSurfDescriptors (Array::Array2dVec3UInt8 *inputNoBorder, Geometry::InterestPointList &pointList, std::string descriptor, int haarSize, int components, int spaciality) |
this code is analogous to CalculateFIST2Descriptors | |
ILOG_VAR_INIT (TestFunctions, Impala.Core.Feature) | |
CPPUNIT_TEST_SUITE_REGISTRATION (TestFunctions) | |
ILOG_CLASS_INIT (TestMakeRandomTree, Impala.Core.Feature) | |
CPPUNIT_TEST_SUITE_REGISTRATION (TestMakeRandomTree) | |
ILOG_CLASS_INIT (TestRandomTree, Impala.Core.Feature) | |
CPPUNIT_TEST_SUITE_REGISTRATION (TestRandomTree) | |
ILOG_VAR_INIT (VisSem, Impala.Core.Feature) | |
Vector::VectorSet< Core::Array::Array2dScalarReal64 > * | Weibull (Core::Array::Array2dVec3UInt8 *im, Real64 sigma) |
Vector::VectorSet< Core::Array::Array2dScalarReal64 > * | WeibullIRGB (Core::Array::Array2dVec3UInt8 *im, Real64 sigma, bool doMu=false, bool doAnderson=false, bool doArjan=false) |
Variables | |
const Quid | cInvalidCodeWord = 0x8000000000000000LL |