boofcv.factory.tracker

Class FactoryTrackerObjectQuad

java.lang.Object

boofcv.factory.tracker.FactoryTrackerObjectQuad

public class FactoryTrackerObjectQuad
extends java.lang.Object

Factory for implementations of TrackerObjectQuad, a high level interface for tracking user specified objects inside video sequences. As usual, the high level interface makes it easier to use these algorithms at the expensive of algorithm specific features.

Constructor Summary

Constructors

Constructor and Description
FactoryTrackerObjectQuad()

Method Summary

Modifier and Type	Method and Description
static <T extends boofcv.struct.image.ImageGray<T>> TrackerObjectQuad<T>	circulant(ConfigCirculantTracker config, java.lang.Class<T> imageType) Creates the Circulant feature tracker.
static <T extends boofcv.struct.image.ImageBase<T>> TrackerObjectQuad<T>	meanShiftComaniciu2003(ConfigComaniciu2003 config, boofcv.struct.image.ImageType<T> imageType) Implementation of mean-shift which matches the histogram and can handle targets composed of multiple colors.
static <T extends boofcv.struct.image.ImageBase<T>> TrackerObjectQuad<T>	meanShiftLikelihood(int maxIterations, int numBins, double maxPixelValue, MeanShiftLikelihoodType modelType, boofcv.struct.image.ImageType<T> imageType) Very basic and very fast implementation of mean-shift which uses a fixed sized rectangle for its region.
static <T extends boofcv.struct.image.ImageGray<T>,D extends boofcv.struct.image.ImageGray<D>> TrackerObjectQuad<T>	sparseFlow(SfotConfig config, java.lang.Class<T> imageType, java.lang.Class<D> derivType) Create an instance of Sparse Flow Object Tracker for the TrackerObjectQuad interface.
static <T extends boofcv.struct.image.ImageGray<T>,D extends boofcv.struct.image.ImageGray<D>> TrackerObjectQuad<T>	tld(ConfigTld config, java.lang.Class<T> imageType) Create an instance of Tracking-Learning-Detection (TLD) tracker for the TrackerObjectQuad interface.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

FactoryTrackerObjectQuad

public FactoryTrackerObjectQuad()

Method Detail

tld

public static <T extends boofcv.struct.image.ImageGray<T>,D extends boofcv.struct.image.ImageGray<D>> TrackerObjectQuad<T> tld(ConfigTld config,
                                                                                                                               java.lang.Class<T> imageType)

Create an instance of Tracking-Learning-Detection (TLD) tracker for the TrackerObjectQuad interface.

Type Parameters:

T - Image input type

D - Image derivative type

Parameters:

config - Configuration for the tracker

Returns:

TrackerObjectQuad

sparseFlow

public static <T extends boofcv.struct.image.ImageGray<T>,D extends boofcv.struct.image.ImageGray<D>> TrackerObjectQuad<T> sparseFlow(SfotConfig config,
                                                                                                                                      java.lang.Class<T> imageType,
                                                                                                                                      java.lang.Class<D> derivType)

Create an instance of Sparse Flow Object Tracker for the TrackerObjectQuad interface.

Type Parameters:

T - Image input type

D - Image derivative type. Null for default.

Parameters:

config - Configuration for the tracker, Null for default.

Returns:

TrackerObjectQuad

meanShiftLikelihood

public static <T extends boofcv.struct.image.ImageBase<T>> TrackerObjectQuad<T> meanShiftLikelihood(int maxIterations,
                                                                                                    int numBins,
                                                                                                    double maxPixelValue,
                                                                                                    MeanShiftLikelihoodType modelType,
                                                                                                    boofcv.struct.image.ImageType<T> imageType)

Very basic and very fast implementation of mean-shift which uses a fixed sized rectangle for its region. Works best when the target is composed of a single color.

Parameters:

maxIterations - Maximum number of mean-shift iterations. Try 30.

numBins - Number of bins in the histogram color model. Try 5.

maxPixelValue - Maximum number of pixel values. For 8-bit images this will be 256

modelType - Type of color model used.

imageType - Type of image

Returns:

TrackerObjectQuad based on TrackerMeanShiftLikelihood.

See Also:

TrackerMeanShiftLikelihood

meanShiftComaniciu2003

public static <T extends boofcv.struct.image.ImageBase<T>> TrackerObjectQuad<T> meanShiftComaniciu2003(ConfigComaniciu2003 config,
                                                                                                       boofcv.struct.image.ImageType<T> imageType)

Implementation of mean-shift which matches the histogram and can handle targets composed of multiple colors. The tracker can also be configured to estimate gradual changes in scale. The track region is composed of a rotated rectangle.

Type Parameters:

T - Image type

Parameters:

config - Tracker configuration

Returns:

TrackerObjectQuad based on Comaniciu2003

See Also:

TrackerMeanShiftComaniciu2003

circulant

public static <T extends boofcv.struct.image.ImageGray<T>> TrackerObjectQuad<T> circulant(ConfigCirculantTracker config,
                                                                                          java.lang.Class<T> imageType)

Creates the Circulant feature tracker. Texture based tracker which uses the theory of circulant matrices, Discrete Fourier Transform (DCF), and linear classifiers to track a target. Fixed sized rectangular target and only estimates translation. Can't detect when it loses track or re-aquire track.

Parameters:

config - Configuration

Returns:

CirculantTracker

See Also:

CirculantTracker