gov.nih.mipav.model.algorithms.AlgorithmColocalizationEM

All Implemented Interfaces:: ActionListener, WindowListener, Runnable, EventListener

public class AlgorithmColocalizationEM extends AlgorithmBase

An optional registration may be performed before colocalization. In this registration both images or colors are moved halfway to a common center point. AlgorithmRegOAR2D is used with the cost function being the only registration parameter the user can vary in the dialog box. Correlation ratio is the default, but the user can also select least squares, normalized cross correlation, or normalized mutual information. The colocalization will be performed on the registered image or images rather than on the original image. If the maximum data range is decreased, the bin1 or bin2 will be decreased if necessary to keep them from exceeding the actual number of channels in the data range.

A histogram buffer is generated with bin1 counts along the x axis and bin2 counts along the y axis. bin1 and bin2 are user selectable in the dialog. leftPad, rightPad, bottomPad, and topPad spaces are left around the data area in histBuffer to allow room for axis lines and labels.

The user specifies the number of gaussian distributions. For each distribution the initial values of the means, standard deviations, and correlation are assigned using a random number generator. Each distribution is initially assigned an equal probability.

The ellipse corresponds to a contour of constant probabliity density on which the density is smaller by a factor of exp[-0.5] than it is at the mean point. The probability density function is a jointly gaussian or bivariate gaussian density function. In each iteration: 1.) Find the gaussian distribution j with the form: p(x,y/j) = (1/(2*PI*stdxj*stdyj*sqrt(1 - rhoj**2))) * exp{(-1/(2*(1 - rhoj**2)))[((x-meanxj)**2)/(stdxj**2) - (2*rhoj*(x-meanxj)*(y-meanyj))/(stdxj*stdyj) + ((y - meanyj)**2)/(stdyj**2)]} where stdxj and stdyj are the x and y standard deviations and rhoj is the correlation coefficient.

2.) p(x,y) are found from p(x,y) = sum over j of p(x,y/j)*p(j). 3.) Pold(j/(x,y) = (sum over all j of Pold(x,y/j)*Pold(j))/Pold(x,y)

The update equations for the values of the new variables for each gaussian distribution j are: 4.) meanxj = (sum over x,y of Pold(j/(x,y))*x)/(sum over x,y of Pold(j/(x,y))) 5.) meanyj = (sum over x,y of Pold(j/(x,y))*y)/(sum over x,y of Pold(j/(x,y))) 6.) stdxj**2 = (sum over x,y of Pold(j/(x,y)) * (x - meanxj)**2)/ (sum over x,y of Pold(j/(x,y))) 7.) stdyj**2 = (sum over x,y of Pold(j/(x,y)) * (y - meanyj)**2)/ (sum over x,y of Pold(j/(x,y))) 8.) rhoj = (sum over x,y of Pold(j/(x,y)) * (x - meanxj) * (y - meanyj)/(stdjx * stdjy))/ (sum over x,y of Pold(j/(x,y)))

9.) Each new p[j] is found from (sum over x,y of Pold(j/(x,y)))/number of (x,y) points These 9 basic steps are iterated the number of times specified by the user.

The calculated data is passed to ViewJFrameColocalization.

References:

1.) Expectation Maximization Algorithm by Ismet Bayraktaroglu (ismet@bayraktaroglu.org) Code for circles was originally found at http://www-cse.ucsd.edu/users/ibayrakt/java/em but this site is no longer available.

2.) Neural Networks for Pattern Recognition by Christopher M. Bishop, Oxford University Press, 1995, Section 2.6.2 The EM algorithm, pp. 65 - 72.

3.) Probability, Random Variables, and Random Signal Principles, Second Edition by Peyton Z. Peebles, Jr., McGraw-Hill Book Company, 1987, Section 5.3 Jointly Gaussian Random Variables, pp. 124 - 128.

4.) Clustering Algorithms by John A. Hartigan, Wiley, John invalid input: '&' Sons, Incorporated, 1975, Chapter 5, Mixtures, pp. 113 - 129.

5.) Principles of Data Mining by David Hand, Heikki Mannila, and Padhraic Smith, A Bradford Book The MIT Press, 2001, pp. 260 - 265, 281 - 284, 315 - 323.

Nested Class Summary

Nested classes/interfaces inherited from class java.lang.Thread
Thread.Builder, Thread.State, Thread.UncaughtExceptionHandler
Field Summary

Fields

Modifier and Type

Field

Description

(package private) float[]

b1

DOCUMENT ME!

(package private) float[]

b2

DOCUMENT ME!

private ModelImage

baseImage

DOCUMENT ME!

private int

bin1

DOCUMENT ME!

private int

bin2

DOCUMENT ME!

private int

bottomPad

DOCUMENT ME!

(package private) float[]

buffer

DOCUMENT ME!

private int

cost

DOCUMENT ME!

private int

count

DOCUMENT ME!

private double[]

dMax

DOCUMENT ME!

private double[]

dMin

DOCUMENT ME!

private boolean

doOr

If true, or thresholds.

private boolean

entireImage

DOCUMENT ME!

private int

gaussians

DOCUMENT ME!

private double[]

halfMajor

DOCUMENT ME!

private double[]

halfMinor

DOCUMENT ME!

(package private) BitSet

inputMask

DOCUMENT ME!

private int

iterations

DOCUMENT ME!

private int

leftPad

Space padding around the histogram data area.

private double[][]

mean

DOCUMENT ME!

private double[][]

meanOld

DOCUMENT ME!

private byte[]

mostProbableClass

DOCUMENT ME!

private double[]

prob

DOCUMENT ME!

private double[]

probOld

DOCUMENT ME!

private double[]

probX

DOCUMENT ME!

private double[]

probXOld

DOCUMENT ME!

private boolean

register

DOCUMENT ME!

private double[]

rho

DOCUMENT ME!

private double[]

rhoOld

DOCUMENT ME!

private int

rightPad

DOCUMENT ME!

(package private) float[]

secondBuffer

DOCUMENT ME!

private byte[]

segBuffer

DOCUMENT ME!

private ModelImage

segImage

Displays classes to which pixels belong.

private double[][]

std

DOCUMENT ME!

private double[][]

stdOld

DOCUMENT ME!

private double[]

theta

DOCUMENT ME!

private float

threshold1

The minimum values that are considered as data values.

private float

threshold2

DOCUMENT ME!

private int

topPad

DOCUMENT ME!

private boolean

useBlue

DOCUMENT ME!

private boolean

useGreen

DOCUMENT ME!

private boolean

useRed

DOCUMENT ME!

(package private) int

xDim

DOCUMENT ME!

(package private) int

yDim

DOCUMENT ME!

(package private) int

zDim

DOCUMENT ME!

Fields inherited from class gov.nih.mipav.model.algorithms.AlgorithmBase
destFlag, destImage, image25D, mask, maxProgressValue, minProgressValue, multiThreadingEnabled, nthreads, progress, progressModulus, progressStep, runningInSeparateThread, separable, srcImage, threadStopped

Fields inherited from class java.lang.Thread
MAX_PRIORITY, MIN_PRIORITY, NORM_PRIORITY
Constructor Summary

Constructors

Constructor

Description

AlgorithmColocalizationEM(ModelImage destImg, ModelImage segImage, ModelImage srcImg, int bin1, int bin2, float threshold1, float threshold2, boolean doOr, int leftPad, int rightPad, int bottomPad, int topPad, boolean useRed, boolean useGreen, boolean useBlue, boolean entireImage, boolean register, int cost, int gaussians, int iterations)

Constructor for images in which 2D histogram is placed in a predetermined destination image.

AlgorithmColocalizationEM(ModelImage destImg, ModelImage segImage, ModelImage srcImg, ModelImage baseImage, int bin1, int bin2, float threshold1, float threshold2, boolean doOr, int leftPad, int rightPad, int bottomPad, int topPad, boolean entireImage, boolean register, int cost, int gaussians, int iterations)

Constructor for images in which 2D histogram is placed in a predetermined destination image.
Method Summary

Modifier and Type

Method

Description

private void

calc2D()

This function produces a 2D histogram image with srcImage values represented across the x axis and baseImage values represented across the y axis.

private void

calc3D()

This function produces a 2D histogram image with srcImage values represented across the x axis and baseImage values represented across the y axis.

private void

calcColor2D()

This function produces a 2D histogram image with first color values represented across the x axis and second color values represented across the y axis.

private void

calcColor3D()

This function produces a 2D histogram image with first color values represented across the x axis and second color values represented across the y axis.

private void

emfunction()

DOCUMENT ME!

void

finalize()

Prepares this class for destruction.

private void

getProbX()

DOCUMENT ME!

private void

getProbXOld()

DOCUMENT ME!

private void

newToOld()

DOCUMENT ME!

private double

probjxOld(int i, int j)

DOCUMENT ME!

private double

probXj(int i, int j)

DOCUMENT ME!

private double

probXjOld(int i, int j)

DOCUMENT ME!

void

runAlgorithm()

starts the algorithm.

void

updateFileInfo(ModelImage image, ModelImage resultImage)

Copy important file information to resultant image structure.

Methods inherited from class gov.nih.mipav.model.algorithms.AlgorithmBase
actionPerformed, addListener, addProgressChangeListener, calculateImageSize, calculatePrincipleAxis, computeElapsedTime, computeElapsedTime, convertIntoFloat, delinkProgressToAlgorithm, delinkProgressToAlgorithmMulti, displayError, errorCleanUp, fireProgressStateChanged, fireProgressStateChanged, fireProgressStateChanged, fireProgressStateChanged, fireProgressStateChanged, generateProgressValues, getDestImage, getElapsedTime, getMask, getMaxProgressValue, getMinProgressValue, getNumberOfThreads, getProgress, getProgressChangeListener, getProgressChangeListeners, getProgressModulus, getProgressStep, getProgressValues, getSrcImage, isCompleted, isImage25D, isMultiThreadingEnabled, isRunningInSeparateThread, isThreadStopped, linkProgressToAlgorithm, linkProgressToAlgorithm, makeProgress, notifyListeners, removeListener, removeProgressChangeListener, run, setCompleted, setImage25D, setMask, setMaxProgressValue, setMinProgressValue, setMultiThreadingEnabled, setNumberOfThreads, setProgress, setProgressModulus, setProgressStep, setProgressValues, setProgressValues, setRunningInSeparateThread, setSrcImage, setStartTime, setThreadStopped, startMethod, windowActivated, windowClosed, windowClosing, windowDeactivated, windowDeiconified, windowIconified, windowOpened

Methods inherited from class java.lang.Thread
activeCount, checkAccess, clone, countStackFrames, currentThread, dumpStack, enumerate, getAllStackTraces, getContextClassLoader, getDefaultUncaughtExceptionHandler, getId, getName, getPriority, getStackTrace, getState, getThreadGroup, getUncaughtExceptionHandler, holdsLock, interrupt, interrupted, isAlive, isDaemon, isInterrupted, isVirtual, join, join, join, join, ofPlatform, ofVirtual, onSpinWait, resume, setContextClassLoader, setDaemon, setDefaultUncaughtExceptionHandler, setName, setPriority, setUncaughtExceptionHandler, sleep, sleep, sleep, start, startVirtualThread, stop, suspend, threadId, toString, yield

Methods inherited from class java.lang.Object
equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Details
- b1
  
  float[] b1
  
  DOCUMENT ME!
- b2
  
  float[] b2
  
  DOCUMENT ME!
- buffer
  
  float[] buffer
  
  DOCUMENT ME!
- inputMask
  
  BitSet inputMask
  
  DOCUMENT ME!
- secondBuffer
  
  float[] secondBuffer
  
  DOCUMENT ME!
- xDim
  
  int xDim
  
  DOCUMENT ME!
- yDim
  
  int yDim
  
  DOCUMENT ME!
- zDim
  
  int zDim
  
  DOCUMENT ME!
- baseImage
  
  private ModelImage baseImage
  
  DOCUMENT ME!
- bin1
  
  private int bin1
  
  DOCUMENT ME!
- bin2
  
  private int bin2
  
  DOCUMENT ME!
- bottomPad
  
  private int bottomPad
  
  DOCUMENT ME!
- cost
  
  private int cost
  
  DOCUMENT ME!
- count
  
  private int count
  
  DOCUMENT ME!
- dMax
  
  private double[] dMax
  
  DOCUMENT ME!
- dMin
  
  private double[] dMin
  
  DOCUMENT ME!
- doOr
  
  private boolean doOr
  
  If true, or thresholds. If false, and thresholds
- entireImage
  
  private boolean entireImage
  
  DOCUMENT ME!
- gaussians
  
  private int gaussians
  
  DOCUMENT ME!
- halfMajor
  
  private double[] halfMajor
  
  DOCUMENT ME!
- halfMinor
  
  private double[] halfMinor
  
  DOCUMENT ME!
- iterations
  
  private int iterations
  
  DOCUMENT ME!
- leftPad
  
  private int leftPad
  
  Space padding around the histogram data area.
- mean
  
  private double[][] mean
  
  DOCUMENT ME!
- meanOld
  
  private double[][] meanOld
  
  DOCUMENT ME!
- mostProbableClass
  
  private byte[] mostProbableClass
  
  DOCUMENT ME!
- prob
  
  private double[] prob
  
  DOCUMENT ME!
- probOld
  
  private double[] probOld
  
  DOCUMENT ME!
- probX
  
  private double[] probX
  
  DOCUMENT ME!
- probXOld
  
  private double[] probXOld
  
  DOCUMENT ME!
- register
  
  private boolean register
  
  DOCUMENT ME!
- rho
  
  private double[] rho
  
  DOCUMENT ME!
- rhoOld
  
  private double[] rhoOld
  
  DOCUMENT ME!
- rightPad
  
  private int rightPad
  
  DOCUMENT ME!
- segBuffer
  
  private byte[] segBuffer
  
  DOCUMENT ME!
- segImage
  
  private ModelImage segImage
  
  Displays classes to which pixels belong.
- std
  
  private double[][] std
  
  DOCUMENT ME!
- stdOld
  
  private double[][] stdOld
  
  DOCUMENT ME!
- theta
  
  private double[] theta
  
  DOCUMENT ME!
- threshold1
  
  private float threshold1
  
  The minimum values that are considered as data values.
- threshold2
  
  private float threshold2
  
  DOCUMENT ME!
- topPad
  
  private int topPad
  
  DOCUMENT ME!
- useBlue
  
  private boolean useBlue
  
  DOCUMENT ME!
- useGreen
  
  private boolean useGreen
  
  DOCUMENT ME!
- useRed
  
  private boolean useRed
  
  DOCUMENT ME!
Constructor Details
- AlgorithmColocalizationEM
  
  public AlgorithmColocalizationEM(ModelImage destImg, ModelImage segImage, ModelImage srcImg, ModelImage baseImage, int bin1, int bin2, float threshold1, float threshold2, boolean doOr, int leftPad, int rightPad, int bottomPad, int topPad, boolean entireImage, boolean register, int cost, int gaussians, int iterations)
  
  Constructor for images in which 2D histogram is placed in a predetermined destination image. Used for 2 black and white images
  
  Parameters:
  
  destImg - image model where result image is to stored
  
  segImage - image to display classes to which pixels belong
  
  srcImg - image for x axis of histogram
  
  baseImage - image for y axis of histogram
  
  bin1 - histogram x axis bin number
  
  bin2 - histogram y axis bin number
  
  threshold1 - minimum value for data in sourceImage
  
  threshold2 - minimum value for data in baseImage
  
  doOr - If true, or thresholds. If false, and thresholds.
  
  leftPad - space to the left of the histogram bin area
  
  rightPad - space to the right of the histogram bin area
  
  bottomPad - space to the bottom of the histogram bin area
  
  topPad - space to the top of the histogram bin area
  
  entireImage - If false, only use data from VOI region
  
  register - If true register 2 black and white images before colocalization
  
  cost - Cost function used in registration
  
  gaussians - Number of gaussians
  
  iterations - Number of iterations
- AlgorithmColocalizationEM
  
  public AlgorithmColocalizationEM(ModelImage destImg, ModelImage segImage, ModelImage srcImg, int bin1, int bin2, float threshold1, float threshold2, boolean doOr, int leftPad, int rightPad, int bottomPad, int topPad, boolean useRed, boolean useGreen, boolean useBlue, boolean entireImage, boolean register, int cost, int gaussians, int iterations)
  
  Constructor for images in which 2D histogram is placed in a predetermined destination image. Used for 1 color image.
  
  Parameters:
  
  destImg - image model where result image is to stored
  
  segImage - image to display classes to which pixels belong
  
  srcImg - source image model
  
  bin1 - histogram x axis bin number
  
  bin2 - histogram y axis bin number
  
  threshold1 - minimum data value for first color
  
  threshold2 - minimum data value for second color
  
  doOr - If true, or thresholds. If false, and thresholds.
  
  leftPad - space to the left of the histogram bin area
  
  rightPad - space to the right of the histogram bin area
  
  bottomPad - space to the bottom of the histogram bin area
  
  topPad - space to the top of the histogram bin area
  
  useRed - DOCUMENT ME!
  
  useGreen - DOCUMENT ME!
  
  useBlue - DOCUMENT ME!
  
  entireImage - If false, use only data from VOI region
  
  register - If true, register 2 colors before colocalization
  
  cost - Cost function used in registration
  
  gaussians - Number of gaussians
  
  iterations - Number of iterations
Method Details
- finalize
  
  public void finalize()
  
  Prepares this class for destruction.
  
  Overrides:
  
  finalize in class AlgorithmBase
- runAlgorithm
  
  public void runAlgorithm()
  
  starts the algorithm.
  
  Specified by:
  
  runAlgorithm in class AlgorithmBase
- updateFileInfo
  
  public void updateFileInfo(ModelImage image, ModelImage resultImage)
  
  Copy important file information to resultant image structure.
  
  Parameters:
  
  image - Source image.
  
  resultImage - Resultant image.
- calc2D
  
  private void calc2D()
  
  This function produces a 2D histogram image with srcImage values represented across the x axis and baseImage values represented across the y axis.
- calc3D
  
  private void calc3D()
  
  This function produces a 2D histogram image with srcImage values represented across the x axis and baseImage values represented across the y axis.
- calcColor2D
  
  private void calcColor2D()
  
  This function produces a 2D histogram image with first color values represented across the x axis and second color values represented across the y axis.
- calcColor3D
  
  private void calcColor3D()
  
  This function produces a 2D histogram image with first color values represented across the x axis and second color values represented across the y axis.
- emfunction
  
  private void emfunction()
  
  DOCUMENT ME!
- getProbX
  
  private void getProbX()
  
  DOCUMENT ME!
- getProbXOld
  
  private void getProbXOld()
  
  DOCUMENT ME!
- newToOld
  
  private void newToOld()
  
  DOCUMENT ME!
- probjxOld
  
  private double probjxOld(int i, int j)
  
  DOCUMENT ME!
  
  Parameters:
  
  i - DOCUMENT ME!
  
  j - DOCUMENT ME!
  
  Returns:
  
  DOCUMENT ME!
- probXj
  
  private double probXj(int i, int j)
  
  DOCUMENT ME!
  
  Parameters:
  
  i - DOCUMENT ME!
  
  j - DOCUMENT ME!
  
  Returns:
  
  DOCUMENT ME!
- probXjOld
  
  private double probXjOld(int i, int j)
  
  DOCUMENT ME!
  
  Parameters:
  
  i - DOCUMENT ME!
  
  j - DOCUMENT ME!
  
  Returns:
  
  DOCUMENT ME!

Class AlgorithmColocalizationEM

Nested Class Summary

Nested classes/interfaces inherited from class java.lang.Thread

Field Summary

Fields inherited from class gov.nih.mipav.model.algorithms.AlgorithmBase

Fields inherited from class java.lang.Thread

Constructor Summary

Method Summary

Methods inherited from class gov.nih.mipav.model.algorithms.AlgorithmBase

Methods inherited from class java.lang.Thread

Methods inherited from class java.lang.Object

Field Details

b1

b2

buffer

inputMask

secondBuffer

xDim

yDim

zDim

baseImage

bin1

bin2

bottomPad

cost

count

dMax

dMin

doOr

entireImage

gaussians

halfMajor

halfMinor

iterations

leftPad

mean

meanOld

mostProbableClass

prob

probOld

probX

probXOld

register

rho

rhoOld

rightPad

segBuffer

segImage

std

stdOld

theta

threshold1

threshold2

topPad

useBlue

useGreen

useRed

Constructor Details

AlgorithmColocalizationEM

AlgorithmColocalizationEM

Method Details

finalize

runAlgorithm

updateFileInfo

calc2D

calc3D

calcColor2D

calcColor3D

emfunction

getProbX

getProbXOld

newToOld

probjxOld

probXj

probXjOld