de.unihalle.informatik.MiToBo.segmentation.regions.mser

Class DetectMSERs

java.lang.Object

de.unihalle.informatik.Alida.operator.ALDOperator

de.unihalle.informatik.MiToBo.core.operator.MTBOperator

de.unihalle.informatik.MiToBo.segmentation.regions.mser.DetectMSERs

All Implemented Interfaces:

de.unihalle.informatik.Alida.datatypes.ALDConfigurationValidator, de.unihalle.informatik.Alida.operator.events.ALDOperatorExecutionProgressEventListener, EventListener

@ALDAOperator(genericExecutionMode=ALL,
              level=APPLICATION)
public class DetectMSERs
extends MTBOperator

Implementation of maximally stable extremal regions based on ImgLib2.

For details about the implementation and its parameters please refer to the corresponding documentation pages in the ImgLib2 Javadoc API.

Please note that currently only gray-scale images are supported. The implementation handles data up to 3D, i.e. x/y/z. Multiple channels and time steps are ignored, i.e. only the first channel and the first time step of an image is taken into account.

Related paper: Matas et al., Robust wide-baseline stereo from maximally stable extremal regions, Image and Vision Computing, vol. 22, no. 10, pp. 761-767, 2004.

Author:

moeller

See Also:

ImgLib2 Javadoc API for MserTree

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	DetectMSERs.Thresholding_Direction Modes for direction of intensity thresholding.

Nested classes/interfaces inherited from class de.unihalle.informatik.Alida.operator.ALDOperator

de.unihalle.informatik.Alida.operator.ALDOperator.HidingMode, de.unihalle.informatik.Alida.operator.ALDOperator.HistoryConstructionMode

Field Summary

Fields

Modifier and Type	Field and Description
private MTBImageByte	binaryMasks (Optinal) stack with binary segmentation masks.
private boolean	createMasks Flag to enable binary mask creation.
private double	delta Difference between threshold levels to compare.
private int	iDepth Depth of input image, i.e. dimension in z or third axis.
private int	iHeight Height of input image, i.e. dimension in y or second axis.
private MTBImage	inImg Input image to process.
private int	iTimeDim Time dimensionality of input image, i.e. dimension in t or time line.
private int	iWidth Width of input image, i.e. dimension in x or first axis.
private long	maxSize Maximum size of regions.
private double	maxVar Maximum instability score, should be in the interval of [0,1].
private double	minDiversity Minimum diversity score, should be in the interval of [0,1].
private long	minSize Minimum size of regions.
private int	mserCount Count of detected MSERs.
private MTBRegionSetInterface	resultMSERs Set of detected MSER regions.
private DetectMSERs.Thresholding_Direction	tDirection Thresholding direction.

Fields inherited from class de.unihalle.informatik.Alida.operator.ALDOperator

completeDAG, name, operatorExecutionEventlistenerList, portHashAccess, verbose, versionProvider

Constructor Summary

Constructors

Constructor and Description
DetectMSERs() Default constructor.

Method Summary

Modifier and Type	Method and Description
private <T extends net.imglib2.type.numeric.RealType<T>> void	extractMSERs(net.imglib2.img.Img<T> img) Method to calculate MSERs from given image.
MTBRegionSetInterface	getMSERs() Returns set of detected MSERs.
int	getNumberOfMSERs() Returns number of detected MSERs.
protected void	operate() This method does the actual work.
void	setCreateBinaryMasks(boolean b) Enable/disable creation of binary masks.
void	setDelta(double d) Set distance between threshold images to be compared.
void	setInputImage(MTBImage img) Set input image to process.
void	setMaximalInstabilityScore(double ms) Set maximal instability score.
void	setMaximalSize(long ms) Set maximal size of valid regions.
void	setMinimalDiversity(double md) Set minimal diversity.
void	setMinimalSize(long ms) Set minimal size of valid regions.
void	setThresholdingDirection(DetectMSERs.Thresholding_Direction td) Set direction of threshold compuations.

Methods inherited from class de.unihalle.informatik.MiToBo.core.operator.MTBOperator

readResolve

Methods inherited from class de.unihalle.informatik.Alida.operator.ALDOperator

addOperatorExecutionProgressEventListener, addParameter, addParameter, addParameterUnconditioned, fieldContained, fireOperatorExecutionProgressEvent, getALDPortHashAccessKey, getConstructionMode, getDocumentation, getHidingMode, getInactiveParameterNames, getInInoutNames, getInInoutNames, getInNames, getInOutNames, getMissingRequiredInputs, getName, getNumParameters, getOutInoutNames, getOutNames, getParameter, getParameterDescriptor, getParameterDescriptorUnconditioned, getParameterNames, getParameterUnconditioned, getSupplementalNames, getVerbose, getVersion, handleOperatorExecutionProgressEvent, hasInOutParameters, hasParameter, isAnnotatedParameter, isConfigured, print, print, print, printInterface, printInterface, readHistory, reinitializeParameterDescriptors, removeOperatorExecutionProgressEventListener, removeParameter, runOp, runOp, runOp, setConstructionMode, setConstructionMode, setConstructionMode, setHidingMode, setName, setParameter, setParameterUnconditioned, setVerbose, toStringVerbose, unconfiguredItems, validate, validateCustom, validateGeneric, writeHistory, writeHistory, writeHistory

Methods inherited from class java.lang.Object

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

Field Detail

inImg

@Parameter(label="Input Image",
           required=true,
           dataIOOrder=0,
           direction=IN,
           description="Input image")
private transient MTBImage inImg

Input image to process.

delta

@Parameter(label="Delta Threshold",
           required=true,
           dataIOOrder=1,
           direction=IN,
           description="Threshold delta for computing instability score.")
private double delta

Difference between threshold levels to compare.

This difference is to be specified as an absolute intensity difference.

minSize

@Parameter(label="Minimum Region Size",
           required=true,
           dataIOOrder=2,
           direction=IN,
           description="Minimum size (in pixels) of accepted MSERs.")
private long minSize

Minimum size of regions.

maxSize

@Parameter(label="Maximum Region Size",
           required=true,
           dataIOOrder=3,
           direction=IN,
           description="Maximum size (in pixels) of accepted MSERs.")
private long maxSize

Maximum size of regions.

maxVar

@Parameter(label="Maximum Instability Score",
           required=true,
           dataIOOrder=4,
           direction=IN,
           description="Maximum instability score of accepted MSERs, should lie between 0 and 1.")
private double maxVar

Maximum instability score, should be in the interval of [0,1].

The score is defined as $s(R_i) = \frac{|R_i \backslash R_{i-\Delta}|}{|R_i|}$.

The values of the score range from a minimum of 0 for a region of size 0 to a maximum of 1. The MSER algorithms discards regions with an instability score exceeding the given maximum instability score. Accordingly, the maximum score should be set to a value within the range of [0,1]. If the score is set to 0, all regions are discarded, if it is set to 1, all regions are kept.

minDiversity

@Parameter(label="Minimal Diversity",
           required=true,
           dataIOOrder=5,
           direction=IN,
           description="Minimal diversity of adjacent accepted MSERs, should lie between 0 and 1.")
private double minDiversity

Minimum diversity score, should be in the interval of [0,1].

The diversity of two regions A and B is defined as $\frac{|B \backslash A|}{|B|}$.

A region A is discarded if the diversity with regard to B is less or equal to the given minimal diversity. The values of the diversity score range from 0 to 1, i.e., the minimal diversity value should be chosen from this range. A value of 1 discards all regions overlapping with their parents, a value of 0 keeps all regions regardless of any existing overlap.

tDirection

@Parameter(label="Thresholding Direction",
           required=true,
           dataIOOrder=6,
           direction=IN,
           description="If to apply thresholds from dark to bright or vice versa.")
private DetectMSERs.Thresholding_Direction tDirection

Thresholding direction.

See Also:

DetectMSERs.Thresholding_Direction

createMasks

@Parameter(label="Create Binary Masks",
           required=false,
           dataIOOrder=1,
           direction=IN,
           description="Turn on mask creation.")
private transient boolean createMasks

Flag to enable binary mask creation.

mserCount

@Parameter(label="Number of MSERs",
           required=true,
           dataIOOrder=0,
           direction=OUT,
           description="Number of MSERs.")
private transient int mserCount

Count of detected MSERs.

resultMSERs

@Parameter(label="Resulting MSER regions",
           required=true,
           dataIOOrder=1,
           direction=OUT,
           description="Result regions.")
private transient MTBRegionSetInterface resultMSERs

Set of detected MSER regions.

binaryMasks

@Parameter(label="Binary Masks of Segmentation",
           dataIOOrder=2,
           direction=OUT,
           description="Binary masks.")
private transient MTBImageByte binaryMasks

(Optinal) stack with binary segmentation masks.

Only created if createMasks is set to true.

iWidth

private int iWidth

Width of input image, i.e. dimension in x or first axis.

iHeight

private int iHeight

Height of input image, i.e. dimension in y or second axis.

iDepth

private int iDepth

Depth of input image, i.e. dimension in z or third axis.

iTimeDim

private int iTimeDim

Time dimensionality of input image, i.e. dimension in t or time line.

Constructor Detail

DetectMSERs

public DetectMSERs()
            throws de.unihalle.informatik.Alida.exceptions.ALDOperatorException

Default constructor.

Throws:

de.unihalle.informatik.Alida.exceptions.ALDOperatorException

Method Detail

setInputImage

public void setInputImage(MTBImage img)

Set input image to process.

Parameters:

img - Image to process.

setDelta

public void setDelta(double d)

Set distance between threshold images to be compared.

Parameters:

d - Distance to be applied.

setMinimalSize

public void setMinimalSize(long ms)

Set minimal size of valid regions.

Regions smaller than the given threshold are discarded.

Parameters:

ms - Minimal region size.

setMaximalSize

public void setMaximalSize(long ms)

Set maximal size of valid regions.

Regions larger than the given threshold are discarded.

Parameters:

ms - Maximal region size.

setMaximalInstabilityScore

public void setMaximalInstabilityScore(double ms)

Set maximal instability score.

Parameters:

ms - Maximal instability score.

setMinimalDiversity

public void setMinimalDiversity(double md)

Set minimal diversity.

Parameters:

ms - Minimal diversity allowed for valid regions.

setThresholdingDirection

public void setThresholdingDirection(DetectMSERs.Thresholding_Direction td)

Set direction of threshold compuations.

Parameters:

ms - Direction of thresholding.

setCreateBinaryMasks

public void setCreateBinaryMasks(boolean b)

Enable/disable creation of binary masks.

Parameters:

b - If true, masks are created, otherwise not.

getNumberOfMSERs

public int getNumberOfMSERs()

Returns number of detected MSERs.

getMSERs

public MTBRegionSetInterface getMSERs()

Returns set of detected MSERs.

Depending on the dimension of the input image either a set of 2D or 3D regions is returned, i.e. the object is either of type MTBRegion2DSet or MTBRegion3DSet.

operate

protected void operate()
                throws de.unihalle.informatik.Alida.exceptions.ALDOperatorException

This method does the actual work.

Specified by:

operate in class de.unihalle.informatik.Alida.operator.ALDOperator

Throws:

de.unihalle.informatik.Alida.exceptions.ALDOperatorException

extractMSERs

private <T extends net.imglib2.type.numeric.RealType<T>> void extractMSERs(net.imglib2.img.Img<T> img)

Method to calculate MSERs from given image.

Parameters:

img - Image to process.