boofcv.alg.tracker.tld

Class TldVarianceFilter<T extends boofcv.struct.image.ImageGray<T>>

java.lang.Object

boofcv.alg.tracker.tld.TldVarianceFilter<T>

public class TldVarianceFilter<T extends boofcv.struct.image.ImageGray<T>>
extends java.lang.Object

Compute the variance for a rectangular region using the integral image. Supports both U8 and F32 input images. For each new image in the sequence a call to setImage(ImageGray) must be done so that it can compute the required integral images. See paper for mathematical details on how the variance is computed using integral images.

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	TldVarianceFilter()
	TldVarianceFilter(java.lang.Class<T> imageType) Constructor which specifies the input image type.

Method Summary

Modifier and Type	Method and Description
boolean	checkVariance(boofcv.struct.ImageRectangle r) Performs variance test at the specified rectangle
protected double	computeVariance(int x0, int y0, int x1, int y1) Computes the variance inside the specified rectangle.
protected double	computeVarianceSafe(int x0, int y0, int x1, int y1) Computes the variance inside the specified rectangle.
double	getThresholdLower()
void	selectThreshold(boofcv.struct.ImageRectangle r) Selects a threshold based on image statistics.
void	setImage(T gray) Sets the input image.
static void	transformSq(boofcv.struct.image.GrayF32 input, boofcv.struct.image.GrayF64 transformed) Integral image of pixel value squared.
static void	transformSq(boofcv.struct.image.GrayU8 input, boofcv.struct.image.GrayS64 transformed) Integral image of pixel value squared.

Methods inherited from class java.lang.Object

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

Constructor Detail

TldVarianceFilter

public TldVarianceFilter(java.lang.Class<T> imageType)

Constructor which specifies the input image type.

Parameters:

imageType - Either GrayU8 or GrayF32

TldVarianceFilter

protected TldVarianceFilter()

Method Detail

setImage

public void setImage(T gray)

Sets the input image. Must be called before other functions/

Parameters:

gray - input image

selectThreshold

public void selectThreshold(boofcv.struct.ImageRectangle r)

Selects a threshold based on image statistics. The paper suggestions 1/2 the variance in the initial patch

checkVariance

public boolean checkVariance(boofcv.struct.ImageRectangle r)

Performs variance test at the specified rectangle

Returns:

true if it passes and false if not

computeVariance

protected double computeVariance(int x0,
                                 int y0,
                                 int x1,
                                 int y1)

Computes the variance inside the specified rectangle. x0 and y0 must be > 0.

Returns:

variance

computeVarianceSafe

protected double computeVarianceSafe(int x0,
                                     int y0,
                                     int x1,
                                     int y1)

Computes the variance inside the specified rectangle.

Returns:

variance

transformSq

public static void transformSq(boofcv.struct.image.GrayU8 input,
                               boofcv.struct.image.GrayS64 transformed)

Integral image of pixel value squared. integer

transformSq

public static void transformSq(boofcv.struct.image.GrayF32 input,
                               boofcv.struct.image.GrayF64 transformed)

Integral image of pixel value squared. floating point

getThresholdLower

public double getThresholdLower()