Package gov.nih.mipav.model.algorithms

Class AlgorithmBrainSurfaceExtractor

java.lang.Object

java.lang.Thread

gov.nih.mipav.model.algorithms.AlgorithmBase

gov.nih.mipav.model.algorithms.AlgorithmBrainSurfaceExtractor

All Implemented Interfaces:

AlgorithmInterface, java.awt.event.ActionListener, java.awt.event.WindowListener, java.lang.Runnable, java.util.EventListener

public class AlgorithmBrainSurfaceExtractor
extends AlgorithmBase
implements AlgorithmInterface

This class provides an implementation of a second method for segmentation of the brain from a 3D MRI, as opposed to the BET algorithm implemented in AlgorithmBrainExtractor. The algorithm is partially based on the paper describing the Brain Surface Extraction (BSE) algorithm:

     Magnetic Resonance Image Tissue Classification Using a Partial Volume Model

     David W. Shattuck et al.

     NeuroImage 13, 856-876 (2001)

     available at http://www.idealibrary.com

 

Version:

1.0 June 3, 2004

Author:

Evan McCreedy

See Also:

JDialogBrainSurfaceExtractor, AlgorithmBrainExtractor

Nested Class Summary

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

java.lang.Thread.State, java.lang.Thread.UncaughtExceptionHandler

Field Summary

Fields

Modifier and Type	Field	Description
private float	brainRegionSize	The size of the region we have identified as the brain.
private static float	brainRegionThresholdRatio	The proportion of the total number of pixels in the volume that a region should be before we decide that it's probably the brain.
private int	closeIterations	The number of closings to perform.
private int	closeKernel	Use the sized circle kernel for the closing.
private float	closeKernelSize	The size of the kernel to use in closings.
private boolean	doSeparable	Whether to use the edge detection algorithm with a separable convolver.
private float	edgeKernel	The size of the edge detection kernel.
private boolean	erosion25D	Whether to erode / dilate image slices independently.
private int	erosionIterations	The number of erosions/dilations to perform.
private boolean	extractPaint	This flag is set when the algorithm is being used only to extract paint from the brain.
private boolean	fillHolesFlag	Whether to make an effort to fill in any interior holes in the final brain mask.
private int	filterIterations	The number of filter passes to make.
private float	filterKernel	The size of the filter kernel to use.
private ModelImage	image	The source image.
private int	imgSize	The size, in voxels, of the volume.
private java.util.BitSet	paintMask	The actual paint bitmap.
private IntVector	regionPointStack	The stack of voxels in the image that we should try to grow to.
private static float	regionPointStackRatio	Used to initialize the stack size when finding the largest segmented region (i.e. the brain).
private ModelImage	resultImage	The result image.
private boolean	showIntermediateImages	Whether to show the intermediate images generated during the BSE algorithm.
private int	sliceSize	The size, in voxels, of a slice of the volume.
private int	xDim	The image size in the x dimension.
private float	xRes	The image resolution in the x dimension.
private int	yDim	The image size in the y dimension.
private float	yRes	The image resolution in the y dimension.
private int	zDim	The image size in the z dimension.
private float	zRes	The image resolution in the z dimension.

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
AlgorithmBrainSurfaceExtractor(ModelImage srcImg, int fIter, float fKernel, float eKernel, boolean ero25D, int eroIter, float cKernel, int cIter, boolean showIntermediate, boolean noHoles, boolean doSep, boolean extractPaint)	Create an extractor for segmenting the brain from a 3D magnetic resonance image.

Method Summary

Modifier and Type	Method	Description
private static ModelImage	addPadding(ModelImage img)	Pad an image by copying the first and last slices 2 more times so that the laplacian algorithm behaves well in edge conditions.
void	algorithmPerformed(AlgorithmBase algorithm)	This method is required if the AlgorithmPerformed interface is implemented.
private float[]	calculateProgressRatios(int filterIterations, boolean doSeparable, boolean fillHolesFlag)	Calculates the progress ratios for different algorithms.
private int[]	calculateProgressValueBoundary(int index, float[] ratios, int min, int max)	Calculates the boundary of progress value which is assigned to the indexth algorithm.
void	extractBrain()	Isolate the brain from the rest of the MR image.
void	finalize()	Prepares this class for destruction.
ModelImage	findLargestRegion(ModelImage maskImage)	Find the largest region within a mask image.
float	getBrainVolume()	Returns the volume of the brain in units (typically millimeters).
java.util.BitSet	getComputedPaintMask()	Method is used when the "extract paint" option is set.
ModelImage	getResultImage()	Get the final result image.
int	markRegion(int[] imgData, int initIndex, int regionId)	Start at a point in an image and grow to all connected points, marking them and counting as we go.
private static ModelImage	removePadding(ModelImage paddedImg)	Remove the padding slices from the beginning and end of an image.
void	runAlgorithm()	Starts the algorithm.

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, join, join, join, onSpinWait, resume, setContextClassLoader, setDaemon, setDefaultUncaughtExceptionHandler, setName, setPriority, setUncaughtExceptionHandler, sleep, sleep, start, stop, suspend, toString, yield

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

brainRegionThresholdRatio

private static final float brainRegionThresholdRatio

The proportion of the total number of pixels in the volume that a region should be before we decide that it's probably the brain.

See Also:

Constant Field Values

regionPointStackRatio

private static final float regionPointStackRatio

Used to initialize the stack size when finding the largest segmented region (i.e. the brain).

See Also:

Constant Field Values

brainRegionSize

private float brainRegionSize

The size of the region we have identified as the brain.

closeIterations

private int closeIterations

The number of closings to perform.

closeKernel

private final int closeKernel

Use the sized circle kernel for the closing.

See Also:

Constant Field Values

closeKernelSize

private float closeKernelSize

The size of the kernel to use in closings.

doSeparable

private boolean doSeparable

Whether to use the edge detection algorithm with a separable convolver.

edgeKernel

private float edgeKernel

The size of the edge detection kernel.

erosion25D

private boolean erosion25D

Whether to erode / dilate image slices independently.

erosionIterations

private int erosionIterations

The number of erosions/dilations to perform.

extractPaint

private boolean extractPaint

This flag is set when the algorithm is being used only to extract paint from the brain.

fillHolesFlag

private boolean fillHolesFlag

Whether to make an effort to fill in any interior holes in the final brain mask.

filterIterations

private int filterIterations

The number of filter passes to make.

filterKernel

private float filterKernel

The size of the filter kernel to use.

image

private ModelImage image

The source image.

imgSize

private int imgSize

The size, in voxels, of the volume.

paintMask

private java.util.BitSet paintMask

The actual paint bitmap.

regionPointStack

private IntVector regionPointStack

The stack of voxels in the image that we should try to grow to. Made a class member to increase speed.

resultImage

private ModelImage resultImage

The result image.

showIntermediateImages

private boolean showIntermediateImages

Whether to show the intermediate images generated during the BSE algorithm.

sliceSize

private int sliceSize

The size, in voxels, of a slice of the volume.

xDim

private int xDim

The image size in the x dimension.

xRes

private float xRes

The image resolution in the x dimension.

yDim

private int yDim

The image size in the y dimension.

yRes

private float yRes

The image resolution in the y dimension.

zDim

private int zDim

The image size in the z dimension.

zRes

private float zRes

The image resolution in the z dimension.

Constructor Detail

AlgorithmBrainSurfaceExtractor

public AlgorithmBrainSurfaceExtractor(ModelImage srcImg,
                                      int fIter,
                                      float fKernel,
                                      float eKernel,
                                      boolean ero25D,
                                      int eroIter,
                                      float cKernel,
                                      int cIter,
                                      boolean showIntermediate,
                                      boolean noHoles,
                                      boolean doSep,
                                      boolean extractPaint)

Create an extractor for segmenting the brain from a 3D magnetic resonance image.

Parameters:

srcImg - the source image. Should be MR image of the brain

fIter - number of passes to make with the filter

fKernel - gaussian kernel size to use when running the filter

eKernel - kernel size to use with edge detection algo

ero25D - whether to use 2.5D processing in erosion / dilation

eroIter - number of erosions / dialations to perform

cKernel - kernel size in mm to be used in closing

cIter - number of closings to perform

showIntermediate - whether to save intermediate images for display

noHoles - whether to fill in all interior holes of the extracted brain

doSep - whether to use separable convolver for edge detection

extractPaint - whether to extract paint from the brain

Method Detail

algorithmPerformed

public void algorithmPerformed(AlgorithmBase algorithm)

This method is required if the AlgorithmPerformed interface is implemented. It is called by the algorithms when it has completed or failed to to complete, so that the dialog can be display the result image and/or clean up.

Specified by:

algorithmPerformed in interface AlgorithmInterface

Parameters:

algorithm - Algorithm that caused the event.

extractBrain

public void extractBrain()

Isolate the brain from the rest of the MR image.

finalize

public void finalize()

Prepares this class for destruction.

Overrides:

finalize in class AlgorithmBase

findLargestRegion

public ModelImage findLargestRegion(ModelImage maskImage)

Find the largest region within a mask image.

Parameters:

maskImage - the original mask image, with values of 0 or 1

Returns:

ModelImage the image containing only the largest connected component from maskImage

getBrainVolume

public final float getBrainVolume()

Returns the volume of the brain in units (typically millimeters).

Returns:

The volume of the brain in voxel units (typically millimeters).

getComputedPaintMask

public java.util.BitSet getComputedPaintMask()

Method is used when the "extract paint" option is set.

Returns:

the paint mask bitset representing the area not removed from the image during the brain extraction

getResultImage

public ModelImage getResultImage()

Get the final result image.

Returns:

the final extracted brain image

markRegion

public int markRegion(int[] imgData,
                      int initIndex,
                      int regionId)

Start at a point in an image and grow to all connected points, marking them and counting as we go.

Parameters:

imgData - the exported 3d image mask

initIndex - the point to start growing the region from

regionId - what to mark where the region is located

Returns:

the number of voxels in the grown region

runAlgorithm

public void runAlgorithm()

Starts the algorithm.

Specified by:

runAlgorithm in class AlgorithmBase

addPadding

private static ModelImage addPadding(ModelImage img)

Pad an image by copying the first and last slices 2 more times so that the laplacian algorithm behaves well in edge conditions.

Parameters:

img - the image to pad

Returns:

the padded image

removePadding

private static ModelImage removePadding(ModelImage paddedImg)

Remove the padding slices from the beginning and end of an image.

Parameters:

paddedImg - the padded image

Returns:

the image without padding

calculateProgressRatios

private float[] calculateProgressRatios(int filterIterations,
                                        boolean doSeparable,
                                        boolean fillHolesFlag)

Calculates the progress ratios for different algorithms.

Parameters:

filterIterations - the iteration number for filter.

doSeparable - flag to indicate whether use separable algorithm or not.

fillHolesFlag - DOCUMENT ME!

Returns:

the progress ratios for different algorithms

calculateProgressValueBoundary

private int[] calculateProgressValueBoundary(int index,
                                             float[] ratios,
                                             int min,
                                             int max)

Calculates the boundary of progress value which is assigned to the indexth algorithm.

Parameters:

index - the algorithm index

ratios - the ratio array assigned to the algorithms.

min - the minimum progress value available to be distributed

max - the maximum progress value available to be distributed

Returns:

the minimum and maximum progress value assigned to the indexth algorithm.