/**
* 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.
*
* @param algorithm Algorithm that caused the event.
*/
public void algorithmPerformed(AlgorithmBase algorithm) {
if (algorithm instanceof AlgorithmNonlocalMeansFilter) {
image.clearMask();
if ((nlMeansFilterAlgo.isCompleted() == true) && (resultImage != null)) {
updateFileInfo(image, resultImage);
resultImage.clearMask();
// The algorithm has completed and produced a new image to be displayed.
try {
new ViewJFrameImage(resultImage, null, new Dimension(610, 200));
} catch (OutOfMemoryError error) {
MipavUtil.displayError("Out of memory: unable to open new frame");
}
} else if (resultImage == null) {
// These next lines set the titles in all frames where the source image is displayed to
// image name so as to indicate that the image is now unlocked!
// The image frames are enabled and then registed to the userinterface.
Vector<ViewImageUpdateInterface> imageFrames = image.getImageFrameVector();
for (int i = 0; i < imageFrames.size(); i++) {
((Frame) (imageFrames.elementAt(i))).setTitle(titles[i]);
((Frame) (imageFrames.elementAt(i))).setEnabled(true);
if (((Frame) (imageFrames.elementAt(i))) != parentFrame) {
userInterface.registerFrame((Frame) (imageFrames.elementAt(i)));
}
}
if (parentFrame != null) {
userInterface.registerFrame(parentFrame);
}
image.notifyImageDisplayListeners(null, true);
} else if (resultImage != null) {
// algorithm failed but result image still has garbage
resultImage.disposeLocal(); // clean up memory
resultImage = null;
}
}
if (algorithm.isCompleted()) {
insertScriptLine();
}
// save the completion status for later
setComplete(algorithm.isCompleted());
nlMeansFilterAlgo.finalize();
nlMeansFilterAlgo = null;
dispose();
}
/**
* Returns the name of an image output by this algorithm, the image returned depends on the
* parameter label given (which can be used to retrieve the image object from the image registry).
*
* @param imageParamName The output image parameter label for which to get the image name.
* @return The image name of the requested output image parameter label.
*/
public String getOutputImageName(final String imageParamName) {
if (imageParamName.equals(AlgorithmParameters.RESULT_IMAGE)) {
if (getResultImage() != null) {
// algo produced a new result image
return getResultImage().getImageName();
} else {
// algo was done in place
return image.getImageName();
}
}
Preferences.debug(
"Unrecognized output image parameter: " + imageParamName + "\n",
Preferences.DEBUG_SCRIPTING);
return null;
}
/** {@inheritDoc} */
protected void setGUIFromParams() {
image = scriptParameters.retrieveInputImage();
userInterface = ViewUserInterface.getReference();
parentFrame = image.getParentFrame();
if (scriptParameters.doOutputNewImage()) {
setDisplayLocNew();
} else {
setDisplayLocReplace();
}
searchWindowSide = scriptParameters.getParams().getInt("search_window_side");
similarityWindowSide = scriptParameters.getParams().getInt("similarity_window_side");
noiseStandardDeviation = scriptParameters.getParams().getFloat("noise_standard_deviation");
degreeOfFiltering = scriptParameters.getParams().getFloat("degree_of_filtering");
doRician = scriptParameters.getParams().getBoolean("do_rician");
image25D = scriptParameters.doProcess3DAs25D();
}
/**
* Use the GUI results to set up the variables needed to run the algorithm.
*
* @return <code>true</code> if parameters set successfully, <code>false</code> otherwise.
*/
private boolean setVariables() {
String tmpStr;
System.gc();
if (replaceImage.isSelected()) {
displayLoc = REPLACE;
} else if (newImage.isSelected()) {
displayLoc = NEW;
}
tmpStr = textSearchWindowSide.getText();
if (testParameter(tmpStr, 5, 101)) {
searchWindowSide = Integer.valueOf(tmpStr).intValue();
} else {
MipavUtil.displayError("Search window side must be between 5 and 101");
textSearchWindowSide.requestFocus();
textSearchWindowSide.selectAll();
return false;
}
if ((searchWindowSide % 2) == 0) {
MipavUtil.displayError("Search window side must be an odd number");
textSearchWindowSide.requestFocus();
textSearchWindowSide.selectAll();
return false;
}
tmpStr = textSimilarityWindowSide.getText();
if (testParameter(tmpStr, 3, 99)) {
similarityWindowSide = Integer.valueOf(tmpStr).intValue();
} else {
MipavUtil.displayError("Similarity window side must be between 3 and 99");
textSimilarityWindowSide.requestFocus();
textSimilarityWindowSide.selectAll();
return false;
}
if ((similarityWindowSide % 2) == 0) {
MipavUtil.displayError("Similarity window side must be an odd number");
textSimilarityWindowSide.requestFocus();
textSimilarityWindowSide.selectAll();
return false;
}
if (similarityWindowSide >= searchWindowSide) {
MipavUtil.displayError("Similarity window side must be less than search window side");
textSimilarityWindowSide.requestFocus();
textSimilarityWindowSide.selectAll();
return false;
}
tmpStr = textNoiseStandardDeviation.getText();
if (testParameter(tmpStr, 0.001, 1000.0)) {
noiseStandardDeviation = Float.valueOf(tmpStr).floatValue();
} else {
MipavUtil.displayError("Radius must be between 0.001 and 1000.0");
textNoiseStandardDeviation.requestFocus();
textNoiseStandardDeviation.selectAll();
return false;
}
doRician = doRicianCheckBox.isSelected();
if (doRician) {
tmpStr = textDegree.getText();
if (testParameter(tmpStr, 1.0, 10.0)) {
degreeOfFiltering = Float.valueOf(tmpStr).floatValue();
} else {
MipavUtil.displayError("Degree of filtering must be between 1.0 and 10.0");
textDegree.requestFocus();
textDegree.selectAll();
}
}
if (image.getNDims() > 2) {
image25D = image25DCheckBox.isSelected();
}
return true;
}
/**
* Initializes the GUI by creating the components, placing them in the dialog, and displaying
* them.
*/
private void init() {
setForeground(Color.black);
setTitle("Nonlocal Means Filter");
JPanel mainPanel;
mainPanel = new JPanel();
mainPanel.setBorder(BorderFactory.createEmptyBorder(3, 3, 3, 3));
mainPanel.setLayout(new GridBagLayout());
GridBagConstraints gbc = new GridBagConstraints();
gbc.gridwidth = 1;
gbc.gridheight = 1;
gbc.anchor = GridBagConstraints.WEST;
gbc.weightx = 1;
gbc.insets = new Insets(3, 3, 3, 3);
gbc.gridx = 0;
gbc.gridy = 0;
gbc.fill = GridBagConstraints.HORIZONTAL;
paramPanel = new JPanel(new GridBagLayout());
paramPanel.setForeground(Color.black);
paramPanel.setBorder(buildTitledBorder("Parameters"));
mainPanel.add(paramPanel, gbc);
GridBagConstraints gbc2 = new GridBagConstraints();
gbc2.gridwidth = 1;
gbc2.gridheight = 1;
gbc2.anchor = GridBagConstraints.WEST;
gbc2.weightx = 1;
gbc2.insets = new Insets(3, 3, 3, 3);
gbc2.gridx = 0;
gbc2.gridy = 0;
gbc2.fill = GridBagConstraints.HORIZONTAL;
labelSearchWindowSide = createLabel("Search window side (odd)");
paramPanel.add(labelSearchWindowSide, gbc2);
gbc2.gridx = 1;
textSearchWindowSide = createTextField("15");
paramPanel.add(textSearchWindowSide, gbc2);
gbc2.gridx = 0;
gbc2.gridy = 1;
labelSimilarityWindowSide = createLabel("Similarity window side (odd) ");
paramPanel.add(labelSimilarityWindowSide, gbc2);
gbc2.gridx = 1;
textSimilarityWindowSide = createTextField("7");
paramPanel.add(textSimilarityWindowSide, gbc2);
gbc2.gridx = 0;
gbc2.gridy = 2;
labelNoiseStandardDeviation = createLabel("Noise standard deviation ");
paramPanel.add(labelNoiseStandardDeviation, gbc2);
gbc2.gridx = 1;
textNoiseStandardDeviation = createTextField("10.0");
paramPanel.add(textNoiseStandardDeviation, gbc2);
gbc2.gridx = 0;
gbc2.gridy = 3;
labelDegree = createLabel("Degree of filtering ");
labelDegree.setEnabled(doRician);
paramPanel.add(labelDegree, gbc2);
gbc2.gridx = 1;
textDegree = createTextField("1.414");
textDegree.setEnabled(doRician);
paramPanel.add(textDegree, gbc2);
gbc2.gridx = 0;
gbc2.gridy = 4;
doRicianCheckBox = new JCheckBox("Deal with Rician noise in MRI");
doRicianCheckBox.setFont(serif12);
doRicianCheckBox.setSelected(false);
doRicianCheckBox.addActionListener(this);
paramPanel.add(doRicianCheckBox, gbc2);
if (image.getNDims() > 2) {
gbc2.gridx = 0;
gbc2.gridy = 5;
gbc2.gridwidth = 2;
image25DCheckBox = new JCheckBox("Process each slice independently (2.5D)");
image25DCheckBox.setFont(serif12);
paramPanel.add(image25DCheckBox, gbc2);
image25DCheckBox.setSelected(false);
} // if (image.getNDims > 2)
JPanel outputOptPanel = new JPanel(new GridLayout(1, 2));
destinationPanel = new JPanel(new BorderLayout());
destinationPanel.setForeground(Color.black);
destinationPanel.setBorder(buildTitledBorder("Destination"));
outputOptPanel.add(destinationPanel);
destinationGroup = new ButtonGroup();
newImage = new JRadioButton("New image", true);
newImage.setBounds(10, 16, 120, 25);
newImage.setFont(serif12);
destinationGroup.add(newImage);
destinationPanel.add(newImage, BorderLayout.NORTH);
replaceImage = new JRadioButton("Replace image", false);
replaceImage.setFont(serif12);
destinationGroup.add(replaceImage);
destinationPanel.add(replaceImage, BorderLayout.CENTER);
// Only if the image is unlocked can it be replaced.
if (image.getLockStatus() == ModelStorageBase.UNLOCKED) {
replaceImage.setEnabled(true);
} else {
replaceImage.setEnabled(false);
}
gbc.gridx = 0;
gbc.gridy = 1;
mainPanel.add(outputOptPanel, gbc);
mainDialogPanel.add(mainPanel, BorderLayout.CENTER);
mainDialogPanel.add(buildButtons(), BorderLayout.SOUTH);
getContentPane().add(mainDialogPanel);
pack();
setResizable(true);
// setVisible(true);
System.gc();
}
/**
* Once all the necessary variables are set, call the Nonlocal Means filter algorithm based on
* what type of image this is and whether or not there is a separate destination image.
*/
protected void callAlgorithm() {
String name = makeImageName(image.getImageName(), "_NonlocalMeans");
int[] destExtents;
if (image.getNDims() == 2) { // source image is 2D
destExtents = new int[2];
destExtents[0] = image.getExtents()[0]; // X dim
destExtents[1] = image.getExtents()[1]; // Y dim
} else {
destExtents = new int[3];
destExtents[0] = image.getExtents()[0];
destExtents[1] = image.getExtents()[1];
destExtents[2] = image.getExtents()[2];
}
if (displayLoc == NEW) {
try {
// Make result image of float type
if (image.isColorImage()) {
resultImage = new ModelImage(ModelImage.ARGB, destExtents, name);
} else {
resultImage = new ModelImage(ModelImage.FLOAT, destExtents, name);
}
// resultImage = (ModelImage)image.clone();
// resultImage.setImageName(name);
// Make algorithm
nlMeansFilterAlgo =
new AlgorithmNonlocalMeansFilter(
resultImage,
image,
searchWindowSide,
similarityWindowSide,
noiseStandardDeviation,
degreeOfFiltering,
doRician,
image25D);
// This is very important. Adding this object as a listener allows the algorithm to
// notify this object when it has completed of failed. See algorithm performed event.
// This is made possible by implementing AlgorithmedPerformed interface
nlMeansFilterAlgo.addListener(this);
createProgressBar(image.getImageName(), nlMeansFilterAlgo);
// Hide dialog
setVisible(false);
if (isRunInSeparateThread()) {
// Start the thread as a low priority because we wish to still have user interface work
// fast
if (nlMeansFilterAlgo.startMethod(Thread.MIN_PRIORITY) == false) {
MipavUtil.displayError("A thread is already running on this object");
}
} else {
nlMeansFilterAlgo.run();
}
} catch (OutOfMemoryError x) {
MipavUtil.displayError("Dialog Nonlocal Means Filter: unable to allocate enough memory");
if (resultImage != null) {
resultImage.disposeLocal(); // Clean up memory of result image
resultImage = null;
}
return;
}
} else {
try {
// No need to make new image space because the user has choosen to replace the source image
// Make the algorithm class
nlMeansFilterAlgo =
new AlgorithmNonlocalMeansFilter(
null,
image,
searchWindowSide,
similarityWindowSide,
noiseStandardDeviation,
degreeOfFiltering,
doRician,
image25D);
// This is very important. Adding this object as a listener allows the algorithm to
// notify this object when it has completed of failed. See algorithm performed event.
// This is made possible by implementing AlgorithmedPerformed interface
nlMeansFilterAlgo.addListener(this);
createProgressBar(image.getImageName(), nlMeansFilterAlgo);
// Hide the dialog since the algorithm is about to run.
setVisible(false);
// These next lines set the titles in all frames where the source image is displayed to
// "locked - " image name so as to indicate that the image is now read/write locked!
// The image frames are disabled and then unregisted from the userinterface until the
// algorithm has completed.
Vector<ViewImageUpdateInterface> imageFrames = image.getImageFrameVector();
titles = new String[imageFrames.size()];
for (int i = 0; i < imageFrames.size(); i++) {
titles[i] = ((Frame) (imageFrames.elementAt(i))).getTitle();
((Frame) (imageFrames.elementAt(i))).setTitle("Locked: " + titles[i]);
((Frame) (imageFrames.elementAt(i))).setEnabled(false);
userInterface.unregisterFrame((Frame) (imageFrames.elementAt(i)));
}
if (isRunInSeparateThread()) {
// Start the thread as a low priority because we wish to still have user interface work
// fast
if (nlMeansFilterAlgo.startMethod(Thread.MIN_PRIORITY) == false) {
MipavUtil.displayError("A thread is already running on this object");
}
} else {
nlMeansFilterAlgo.run();
}
} catch (OutOfMemoryError x) {
MipavUtil.displayError("Dialog Nonlocal Means Filter: unable to allocate enough memory");
return;
}
}
}