Class TrackerKCF
KCF is a novel tracking framework that utilizes properties of circulant matrix to enhance the processing speed. The original paper of KCF is available at http://home.isr.uc.pt/~henriques/circulant/index.html as well as the matlab implementation.
public class TrackerKCF : Tracker, IDisposable
- Inheritance
-
TrackerKCF
- Implements
- Inherited Members
Constructors
TrackerKCF(float, float, float, float, float, float, bool, bool, bool, bool, int, int, Mode, Mode)
Creates a KCF Tracker
public TrackerKCF(float detectThresh = 0.5, float sigma = 0.2, float lambda = 0.01, float interpFactor = 0.075, float outputSigmaFactor = 0.0625, float pcaLearningRate = 0.15, bool resize = true, bool splitCoeff = true, bool wrapKernel = false, bool compressFeature = true, int maxPatchSize = 6400, int compressedSize = 2, TrackerKCF.Mode descPca = Mode.Cn, TrackerKCF.Mode descNpca = Mode.Gray)
Parameters
detectThresh
floatdetection confidence threshold
sigma
floatgaussian kernel bandwidth
lambda
floatregularization
interpFactor
floatlinear interpolation factor for adaptation
outputSigmaFactor
floatspatial bandwidth (proportional to target)
pcaLearningRate
floatcompression learning rate
resize
boolactivate the resize feature to improve the processing speed
splitCoeff
boolsplit the training coefficients into two matrices
wrapKernel
boolwrap around the kernel values
compressFeature
boolactivate the pca method to compress the features
maxPatchSize
intthreshold for the ROI size
compressedSize
intfeature size after compression
descPca
TrackerKCF.Modecompressed descriptors of TrackerKCF::MODE
descNpca
TrackerKCF.Modenon-compressed descriptors of TrackerKCF::MODE
Methods
DisposeObject()
Release the unmanaged resources associated with this tracker
protected override void DisposeObject()