Skip to content

Instantly share code, notes, and snippets.

@Svidro
Last active May 16, 2019 15:51
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save Svidro/eaf29399e36d34caacddf345ab668b0c to your computer and use it in GitHub Desktop.
Save Svidro/eaf29399e36d34caacddf345ab668b0c to your computer and use it in GitHub Desktop.
Edited Java files
Files from QuPath that can be swapped in place of their default counterparts using Gradle for alternate functionality.
Now adding modifications to the UI
TOC
AbstractTileableDetectionPlugin.java - Changes the tiling size to work better for high resolution images where the tiles might break up
too many individual cells.
Add button.groovy - script that adds a button to the UI that... can do things. If you make it do them.
WatershedCellDetection.java - Gives users the ability to choose weights for various stains when detecting cells. This is in addition to
"OD sum" and "Hematoxylin stain" option in the dropdown menu.
//Found at qupath-core-processing-awt/src/main/java/qupath/lib/plugins/AbstractTileableDetectionPlugin.java
//swaps to a much larger tile size for segmentation of very high resolution images
/*-
* #%L
* This file is part of QuPath.
* %%
* Copyright (C) 2014 - 2016 The Queen's University of Belfast, Northern Ireland
* Contact: IP Management (ipmanagement@qub.ac.uk)
* %%
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as
* published by the Free Software Foundation, either version 3 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public
* License along with this program. If not, see
* <http://www.gnu.org/licenses/gpl-3.0.html>.
* #L%
*/
package qupath.lib.plugins;
import java.util.ArrayList;
import java.util.Collection;
import java.util.List;
import java.util.concurrent.atomic.AtomicInteger;
import qupath.lib.geom.ImmutableDimension;
import qupath.lib.images.ImageData;
import qupath.lib.images.servers.ServerTools;
import qupath.lib.objects.PathObject;
import qupath.lib.plugins.parameters.ParameterList;
import qupath.lib.roi.PathROIToolsAwt;
import qupath.lib.roi.interfaces.ROI;
/**
* Abstract plugin used for detection tasks that support breaking large regions into smaller ones,
* and analyzing these in parallel - optionally with overlaps.
*
* Particularly useful for tasks such as cell detection.
*
* @author Pete Bankhead
*
* @param <T>
*/
public abstract class AbstractTileableDetectionPlugin<T> extends AbstractDetectionPlugin<T> {
/**
* Get the preferred pixel size that would be used for the specified ImageData and ParameterList.
* This is useful in deciding whether to break large regions into smaller, parallelizable tiles.
*
* @param imageData
* @param params
* @return
*/
protected abstract double getPreferredPixelSizeMicrons(ImageData<T> imageData, ParameterList params);
/**
* Create a new ObjectDetector, compatible with the specified ImageData and ParameterList.
*
* @param imageData
* @param params
* @return
*/
protected abstract ObjectDetector<T> createDetector(final ImageData<T> imageData, final ParameterList params);
/**
* Get an appropriate overlap, in pixels, if analysis of the specified ImageData will be tiled.
*
* If the overlap is 0, then tile boundaries are likely to be visible in the results.
*
* If the overlap is > 0, then the overlap should also be > the expected largest size of a detected object -
* otherwise objects may be lost of trimmed when overlaps are resolved. This is because (currently)
* the resolution of overlapping detections involves taking the largest one, rather than (for example) merging them.
*
* (Merging may be permitted in later versions, but only where measurements are not made by the plugin -
* since merged objects may require different measurements, e.g. for area or mean than can be easily computed
* in a general way from the individual objects being merged).
*
* @param imageData
* @param params
* @return The overlap size in pixels, or 0 if overlapped tiles are not supported.
*/
protected abstract int getTileOverlap(final ImageData<T> imageData, final ParameterList params);
/**
* Intercepts the 'standard' addRunnableTasks to (if necessary) insert ParallelTileObjects along the way,
* thereby breaking an excessively-large parentObject into more manageable pieces.
*
* TODO: Avoid hard-coding what is considered a 'manageable size' or a preferred size for parallel tiles.
*
*/
@Override
protected void addRunnableTasks(ImageData<T> imageData, PathObject parentObject, List<Runnable> tasks) {
// if (detector == null || detector.pathROI != parentObject.getROI())
// detector = new CellDetector();
if (imageData == null)
return;
ParameterList params = getParameterList(imageData);
// Determine appropriate sizes - get a downsample factor that is a power of 2
double downsampleFactor = ServerTools.getDownsampleFactor(imageData.getServer(), getPreferredPixelSizeMicrons(imageData, params), true);
int preferred = (int)(4096 * downsampleFactor);
// int preferred = (int)(2048 * downsampleFactor);
// int preferred = (int)(1536 * downsampleFactor);
// int max = (int)(4096 * downsampleFactor);
int max = (int)(6144 * downsampleFactor);
// int max = (int)(2048 * downsampleFactor);
ImmutableDimension sizePreferred = new ImmutableDimension(preferred, preferred);
ImmutableDimension sizeMax = new ImmutableDimension(max, max);
parentObject.clearPathObjects();
// No tasks to complete
Collection<? extends ROI> pathROIs = PathROIToolsAwt.computeTiledROIs(imageData, parentObject, sizePreferred, sizeMax, false, getTileOverlap(imageData, params));
if (pathROIs.isEmpty())
return;
// Exactly one task to complete
if (pathROIs.size() == 1 && pathROIs.iterator().next() == parentObject.getROI()) {
tasks.add(DetectionPluginTools.createRunnableTask(createDetector(imageData, params), getParameterList(imageData), imageData, parentObject));
return;
}
List<ParallelTileObject> tileList = new ArrayList<>();
AtomicInteger countdown = new AtomicInteger(pathROIs.size());
for (ROI pathROI : pathROIs) {
ParallelTileObject tile = new ParallelTileObject(pathROI, imageData.getHierarchy(), countdown);
parentObject.addPathObject(tile);
for (ParallelTileObject tileTemp : tileList) {
if (tileTemp.suggestNeighbor(tile))
tile.suggestNeighbor(tileTemp);
}
tileList.add(tile);
tasks.add(DetectionPluginTools.createRunnableTask(createDetector(imageData, params), params, imageData, tile));
}
imageData.getHierarchy().fireHierarchyChangedEvent(this);
}
}
//Script from komzy here: https://forum.image.sc/t/button-setonaction-does-not-print-anything-when-scripting/25443
guiscript=true
//import javafx.application.Platform
import javafx.scene.control.Button
import javafx.scene.control.Tooltip
import qupath.lib.gui.QuPathGUI
def qupath = QuPathGUI.getInstance()
def button = new Button('CT')
button.setPrefSize(40, QuPathGUI.iconSize)
button.setTooltip(new Tooltip("Cursor Tracker"));
button.setOnAction {
print("Test Print # 2: Button Clicked")
}
qupath.addToolbarButton(button);
print("Test Print # 1")
//Found at qupath-processing-ij/src/main/java/qupath/imagej/detect/nuclei/WatershedCellDetection.java
/*-
* #%L
* This file is part of QuPath.
* %%
* Copyright (C) 2014 - 2016 The Queen's University of Belfast, Northern Ireland
* Contact: IP Management (ipmanagement@qub.ac.uk)
* %%
* This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU General Public License as
* published by the Free Software Foundation, either version 3 of the
* License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public
* License along with this program. If not, see
* <http://www.gnu.org/licenses/gpl-3.0.html>.
* #L%
*/
package qupath.imagej.detect.nuclei;
import java.awt.image.BufferedImage;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collection;
import java.util.Iterator;
import java.util.LinkedHashMap;
import java.util.List;
import java.util.Map;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import ij.IJ;
import ij.ImagePlus;
import ij.Prefs;
import ij.gui.PolygonRoi;
import ij.gui.Roi;
import ij.gui.Wand;
import ij.measure.Calibration;
import ij.measure.Measurements;
import ij.plugin.filter.EDM;
import ij.plugin.filter.RankFilters;
import ij.process.Blitter;
import ij.process.ByteProcessor;
import ij.process.ColorProcessor;
import ij.process.FloatPolygon;
import ij.process.FloatProcessor;
import ij.process.ImageProcessor;
import ij.process.ImageStatistics;
import ij.process.ShortProcessor;
import qupath.imagej.color.ColorDeconvolutionIJ;
import qupath.imagej.helpers.IJTools;
import qupath.imagej.objects.PathImagePlus;
import qupath.imagej.objects.ROIConverterIJ;
import qupath.imagej.objects.measure.ObjectMeasurements;
import qupath.imagej.processing.MorphologicalReconstruction;
import qupath.imagej.processing.ROILabeling;
import qupath.imagej.processing.RegionalExtrema;
import qupath.imagej.processing.SimpleThresholding;
import qupath.imagej.processing.Watershed;
import qupath.imagej.wrappers.PixelImageIJ;
import qupath.lib.analysis.algorithms.SimpleImage;
import qupath.lib.analysis.stats.RunningStatistics;
import qupath.lib.analysis.stats.StatisticsHelper;
import qupath.lib.color.ColorDeconvolutionStains;
import qupath.lib.images.ImageData;
import qupath.lib.images.PathImage;
import qupath.lib.images.servers.ImageServer;
import qupath.lib.images.servers.ServerTools;
import qupath.lib.measurements.MeasurementListFactory;
import qupath.lib.measurements.MeasurementList;
import qupath.lib.objects.PathCellObject;
import qupath.lib.objects.PathDetectionObject;
import qupath.lib.objects.PathObject;
import qupath.lib.objects.classes.PathClassFactory;
import qupath.lib.objects.helpers.PathObjectTools;
import qupath.lib.plugins.AbstractTileableDetectionPlugin;
import qupath.lib.plugins.ObjectDetector;
import qupath.lib.plugins.parameters.DoubleParameter;
import qupath.lib.plugins.parameters.Parameter;
import qupath.lib.plugins.parameters.ParameterList;
import qupath.lib.roi.PolygonROI;
import qupath.lib.roi.experimental.ShapeSimplifier;
import qupath.lib.roi.interfaces.PathArea;
import qupath.lib.roi.interfaces.ROI;
/**
* Default command for cell detection within QuPath.
*
* Assumes either a nuclear or cytoplasmic staining.
*
* Quantification of membranous staining requires a separate detection command.
*
* @author Pete Bankhead
*
*/
public class WatershedCellDetection extends AbstractTileableDetectionPlugin<BufferedImage> {
private static String[] micronParameters = {
"requestedPixelSizeMicrons",
"backgroundRadiusMicrons",
"medianRadiusMicrons",
"sigmaMicrons",
"minAreaMicrons",
"maxAreaMicrons",
"cellExpansionMicrons",
};
private static String[] pixelParameters = {
// "requestedPixelSize",
"backgroundRadius",
"medianRadius",
"sigma",
"minArea",
"maxArea",
"cellExpansion",
};
private static String[] fluorescenceParameters = {
"detectionImageFluorescence"
};
private static String[] brightfieldParameters = {
"detectionImageBrightfield",
"maxBackground"
};
transient private CellDetector detector;
private final static Logger logger = LoggerFactory.getLogger(WatershedCellDetection.class);
static String IMAGE_OPTICAL_DENSITY = "Optical density sum";
static String IMAGE_HEMATOXYLIN = "Hematoxylin OD";
static String IMAGE_MERGED_STAINS = "Weighted Stains";
ParameterList params;
public WatershedCellDetection() {
params = new ParameterList();
// TODO: Use a better way to determining if pixel size is available in microns
// params.addEmptyParameter("detectionParameters", "Detection parameters", true);
String microns = IJ.micronSymbol + "m";
params.addTitleParameter("Setup parameters");
params.addIntParameter("detectionImageFluorescence", "Choose detection channel", 1, null, "Choose the channel number containing a nucleus counterstain (e.g. DAPI)");
params.addChoiceParameter("detectionImageBrightfield", "Choose detection image", IMAGE_HEMATOXYLIN, Arrays.asList(IMAGE_HEMATOXYLIN, IMAGE_OPTICAL_DENSITY, IMAGE_MERGED_STAINS),
"Transformed image to which to apply the detection");
params.addDoubleParameter("requestedPixelSizeMicrons", "Requested pixel size", .5, microns,
"Choose pixel size at which detection will be performed - higher values are likely to be faster, but may be less accurate; set <= 0 to use the full image resolution");
// params.addDoubleParameter("requestedPixelSize", "Requested downsample factor", 1, "");
params.addTitleParameter("Only used for Weighted Stains selection");
params.addDoubleParameter("stainWeight", "Choose Stain1 Weight", 80, null, "Choose a % multiplier to weight the use of stain 1 in nuclei detection");
params.addDoubleParameter("stainWeight2", "Choose Stain2 Weight", 20, null, "Choose a % multiplier to weight the use of stain 2 in nuclei detection");
params.addDoubleParameter("stainWeight3", "Choose Stain3 Weight", 0, null, "Choose a % multiplier to weight the use of stain 3 in nuclei detection, this should usually be zero");
params.addTitleParameter("Nucleus parameters");
params.addDoubleParameter("backgroundRadiusMicrons", "Background radius", 8, microns,
"Radius for background estimation, should be > the largest nucleus radius, or <= 0 to turn off background subtraction");
params.addDoubleParameter("medianRadiusMicrons", "Median filter radius", 0, microns,
"Radius of median filter used to reduce image texture (optional)");
params.addDoubleParameter("sigmaMicrons", "Sigma", 1.5, microns,
"Sigma value for Gaussian filter used to reduce noise; increasing the value stops nuclei being fragmented, but may reduce the accuracy of boundaries");
params.addDoubleParameter("minAreaMicrons", "Minimum area", 10, microns+"^2",
"Detected nuclei with an area < minimum area will be discarded");
params.addDoubleParameter("maxAreaMicrons", "Maximum area", 400, microns+"^2",
"Detected nuclei with an area > maximum area will be discarded");
params.addDoubleParameter("backgroundRadius", "Background radius", 15, "px",
"Radius for background estimation, should be > the largest nucleus radius, or <= 0 to turn off background subtraction");
params.addDoubleParameter("medianRadius", "Median filter radius", 0, "px",
"Radius of median filter used to reduce image texture (optional)");
params.addDoubleParameter("sigma", "Sigma", 3, "px",
"Sigma value for Gaussian filter used to reduce noise; increasing the value stops nuclei being fragmented, but may reduce the accuracy of boundaries");
params.addDoubleParameter("minArea", "Minimum area", 10, "px^2",
"Detected nuclei with an area < minimum area will be discarded");
params.addDoubleParameter("maxArea", "Maximum area", 1000, "px^2",
"Detected nuclei with an area > maximum area will be discarded");
params.addTitleParameter("Intensity parameters");
params.addDoubleParameter("threshold", "Threshold", 0.1, null,
"Intensity threshold - detected nuclei must have a mean intensity >= threshold");
// params.addDoubleParameter("threshold", "Threshold", 0.1, null, 0, 2.5,
// "Intensity threshold - detected nuclei must have a mean intensity >= threshold");
params.addDoubleParameter("maxBackground", "Max background intensity", 2, null,
"If background radius > 0, detected nuclei occurring on a background > max background intensity will be discarded");
// params.addBooleanParameter("mergeAll", "Merge all", true);
params.addBooleanParameter("watershedPostProcess", "Split by shape", true,
"Split merged detected nuclei based on shape ('roundness')");
params.addBooleanParameter("excludeDAB", "Exclude DAB (membrane staining)", false,
"Set to 'true' if regions of high DAB staining should not be considered nuclei; useful if DAB stains cell membranes");
params.addTitleParameter("Cell parameters");
params.addDoubleParameter("cellExpansionMicrons", "Cell expansion", 5, microns, 0, 25,
"Amount by which to expand detected nuclei to approximate the full cell area");
params.addDoubleParameter("cellExpansion", "Cell expansion", 5, "px",
"Amount by which to expand detected nuclei to approximate the full cell area");
// params.addBooleanParameter("limitExpansionByNucleusSize", "Limit cell expansion by nucleus size", false, "If checked, nuclei will not be expanded by more than their (estimated) smallest diameter in any direction - may give more realistic results for smaller, or 'thinner' nuclei");
params.addBooleanParameter("includeNuclei", "Include cell nucleus", true,
"If cell expansion is used, optionally include/exclude the nuclei within the detected cells");
params.addTitleParameter("General parameters");
params.addBooleanParameter("smoothBoundaries", "Smooth boundaries", true,
"Smooth the detected nucleus/cell boundaries");
params.addBooleanParameter("makeMeasurements", "Make measurements", true,
"Add default shape & intensity measurements during detection");
}
static class CellDetector implements ObjectDetector<BufferedImage> {
private String lastServerPath = null;
//private PathImage<ImagePlus> pathImage; // Caching these cause out of memory errors...
private ROI pathROI;
private List<PathObject> pathObjects = null;
// private WatershedCellDetector detector2;
// private FloatProcessor fpDetection, fpH, fpDAB;
// private ColorDeconvolutionStains stains;
private boolean nucleiClassified = false;
public static double getPreferredPixelSizeMicrons(ImageData<BufferedImage> imageData, ParameterList params) {
if (imageData.getServer().hasPixelSizeMicrons())
return Math.max(params.getDoubleParameterValue("requestedPixelSizeMicrons"), imageData.getServer().getAveragedPixelSizeMicrons());
return Double.NaN;
}
@Override
public Collection<PathObject> runDetection(final ImageData<BufferedImage> imageData, ParameterList params, ROI pathROI) {
// TODO: Give a sensible error
if (pathROI == null)
return null;
// Get a PathImage if we have a new ROI
// boolean imageChanged = false;
PathImage<ImagePlus> pathImage = null;
if (lastServerPath == null || !lastServerPath.equals(imageData.getServerPath()) || pathImage == null || !pathROI.equals(this.pathROI)) {
ImageServer<BufferedImage> server = imageData.getServer();
lastServerPath = imageData.getServerPath();
pathImage = PathImagePlus.createPathImage(server, pathROI, ServerTools.getDownsampleFactor(server, getPreferredPixelSizeMicrons(imageData, params), true));
logger.trace("Cell detection with downsample: " + pathImage.getDownsampleFactor());
this.pathROI = pathROI;
// imageChanged = true;
}
// Create a detector if we don't already have one for this image
boolean isBrightfield = imageData.isBrightfield();
// if (detector2 == null || imageChanged || stains != imageData.getColorDeconvolutionStains()) {
// if (imageChanged || stains != imageData.getColorDeconvolutionStains()) {
ImageProcessor ip = pathImage.getImage().getProcessor();
FloatProcessor fpDetection = null;
ColorDeconvolutionStains stains = imageData.getColorDeconvolutionStains();
Map<String, FloatProcessor> channels = new LinkedHashMap<>();
Map<String, FloatProcessor> channelsCell = new LinkedHashMap<>();
Roi roi = null;
if (pathROI != null)
roi = ROIConverterIJ.convertToIJRoi(pathROI, pathImage);
if (ip instanceof ColorProcessor && stains != null && isBrightfield) {
FloatProcessor[] fps = ColorDeconvolutionIJ.colorDeconvolve((ColorProcessor)ip, stains.getStain(1), stains.getStain(2), stains.getStain(3));
if (params.getChoiceParameterValue("detectionImageBrightfield").equals(IMAGE_MERGED_STAINS)) {
double weight = params.getDoubleParameterValue("stainWeight");
double weight2 = params.getDoubleParameterValue("stainWeight2");
double weight3 = params.getDoubleParameterValue("stainWeight3");
double total_weight = weight+weight2+weight3;
float[][] Chan2 = fps[1].getFloatArray();
float[][] Chan1 = fps[0].getFloatArray();
float[][] Chan3 = fps[2].getFloatArray();
for(int i = 0; i < Chan1.length; i++) {
for(int j = 0; j< Chan1[0].length; j++) {
Chan1[i][j] = (float) (Chan1[i][j]*weight/total_weight + Chan2[i][j]*(weight2)/total_weight+Chan3[i][j]*(weight3)/total_weight);
} // end j for loop
} // end i for loop
FloatProcessor a = new FloatProcessor(Chan1);
channels.put("HDAB OD", a);
channels.put("DAB OD", fps[1]);
channelsCell.put("DAB OD", fps[1]);
fpDetection = (FloatProcessor)a.duplicate();
}else if (stains.isH_DAB()) {
channels.put("Hematoxylin OD", fps[0]);
channels.put("DAB OD", fps[1]);
channelsCell.put("DAB OD", fps[1]);
}else if (stains.isH_E()) {
channels.put("Hematoxylin OD", fps[0]);
channels.put("Eosin OD", fps[1]);
channelsCell.put("Eosin OD", fps[1]);
}
if (!params.getParameters().get("detectionImageBrightfield").isHidden()) {
if (params.getChoiceParameterValue("detectionImageBrightfield").equals(IMAGE_OPTICAL_DENSITY))
fpDetection = ColorDeconvolutionIJ.convertToOpticalDensitySum((ColorProcessor)ip, stains.getMaxRed(), stains.getMaxGreen(), stains.getMaxBlue());
else if (fpDetection == null)
fpDetection = (FloatProcessor)fps[0].duplicate();
}
// Temporary test of the usefulness of RGB measurements...
// channels.put("Red", ((ColorProcessor)ip).toFloat(0, null));
// channels.put("Green", ((ColorProcessor)ip).toFloat(1, null));
// channels.put("Blue", ((ColorProcessor)ip).toFloat(2, null));
} //else {
if (fpDetection == null) {
if (ip instanceof ColorProcessor) {
channels.put("Channel 1", ((ColorProcessor)ip).toFloat(0, null));
channels.put("Channel 2", ((ColorProcessor)ip).toFloat(1, null));
channels.put("Channel 3", ((ColorProcessor)ip).toFloat(2, null));
} else {
ImagePlus imp = pathImage.getImage();
for (int c = 1; c <= imp.getNChannels(); c++) {
channels.put("Channel " + c, imp.getStack().getProcessor(imp.getStackIndex(c, 0, 0)).convertToFloatProcessor());
}
}
// For fluorescence, measure everything
channelsCell.putAll(channels);
// TODO: Deal with fluorescence... for now, defaults to first channel (may be totally wrong)
int detectionChannel = 3;
if (!isBrightfield)
detectionChannel = params.getIntParameterValue("detectionImageFluorescence");
fpDetection = channels.get("Channel " + detectionChannel);
if (fpDetection == null) {
logger.warn("Unable to find specified Channel {} - will default to Channel 3", detectionChannel);
fpDetection = channels.get("Channel 3");
}
}
WatershedCellDetector detector2 = new WatershedCellDetector(fpDetection, channels, channelsCell, roi, pathImage);
// Create or reset the PathObjects list
if (pathObjects == null)
pathObjects = new ArrayList<>();
else
pathObjects.clear();
// Convert parameters where needed
double sigma, medianRadius, backgroundRadius, minArea, maxArea, cellExpansion;
if (pathImage.hasPixelSizeMicrons()) {
double pixelSize = 0.5 * (pathImage.getPixelHeightMicrons() + pathImage.getPixelWidthMicrons());
backgroundRadius = params.getDoubleParameterValue("backgroundRadiusMicrons") / pixelSize;
medianRadius = params.getDoubleParameterValue("medianRadiusMicrons") / pixelSize;
sigma = params.getDoubleParameterValue("sigmaMicrons") / pixelSize;
minArea = params.getDoubleParameterValue("minAreaMicrons") / (pixelSize * pixelSize);
maxArea = params.getDoubleParameterValue("maxAreaMicrons") / (pixelSize * pixelSize);
cellExpansion = params.getDoubleParameterValue("cellExpansionMicrons") / (pixelSize);
} else {
backgroundRadius = params.getDoubleParameterValue("backgroundRadius");
medianRadius = params.getDoubleParameterValue("medianRadius");
sigma = params.getDoubleParameterValue("sigma");
minArea = params.getDoubleParameterValue("minArea");
maxArea = params.getDoubleParameterValue("maxArea");
cellExpansion = params.getDoubleParameterValue("cellExpansion");
}
detector2.runDetection(
backgroundRadius,
isBrightfield ? params.getDoubleParameterValue("maxBackground") : Double.NEGATIVE_INFINITY,
medianRadius,
sigma,
params.getDoubleParameterValue("threshold"),
minArea,
maxArea,
true, // always use 'merge all' params.getBooleanParameterValue("mergeAll"),
params.getBooleanParameterValue("watershedPostProcess"),
params.getBooleanParameterValue("excludeDAB"),
cellExpansion,
// params.getBooleanParameterValue("limitExpansionByNucleusSize"),
params.getBooleanParameterValue("smoothBoundaries"),
params.getBooleanParameterValue("includeNuclei"),
params.getBooleanParameterValue("makeMeasurements"),
pathROI.getZ(),
pathROI.getT());// && isBrightfield);
pathObjects.addAll(detector2.getPathObjects());
return pathObjects;
}
@Override
public String getLastResultsDescription() {
if (pathObjects == null)
return null;
int nDetections = pathObjects.size();
if (nDetections == 1)
return "1 nucleus detected";
String s = String.format("%d nuclei detected", nDetections);
if (nucleiClassified) {
int nPositive = PathObjectTools.countObjectsWithClass(pathObjects, PathClassFactory.getPathClass(PathClassFactory.getPositiveClassName()), false);
int nNegative = PathObjectTools.countObjectsWithClass(pathObjects, PathClassFactory.getPathClass(PathClassFactory.getNegativeClassName()), false);
return String.format("%s (%.3f%% positive)", s, ((double)nPositive * 100.0 / (nPositive + nNegative)));
} else
return s;
}
}
@Override
public ParameterList getDefaultParameterList(final ImageData<BufferedImage> imageData) {
// Show/hide parameters depending on whether the pixel size is known
Map<String, Parameter<?>> map = params.getParameters();
boolean pixelSizeKnown = imageData.getServer() != null && imageData.getServer().hasPixelSizeMicrons();
for (String name : micronParameters)
map.get(name).setHidden(!pixelSizeKnown);
for (String name : pixelParameters)
map.get(name).setHidden(pixelSizeKnown);
params.setHiddenParameters(!pixelSizeKnown, micronParameters);
params.setHiddenParameters(pixelSizeKnown, pixelParameters);
boolean isBrightfield = imageData.isBrightfield();
params.setHiddenParameters(!isBrightfield, brightfieldParameters);
params.setHiddenParameters(isBrightfield, fluorescenceParameters);
if (!isBrightfield) {
if (imageData.getServer().getBitsPerPixel() > 8)
((DoubleParameter)params.getParameters().get("threshold")).setValue(100.0);
else
((DoubleParameter)params.getParameters().get("threshold")).setValue(25.0);
}
// map.get("detectionImageBrightfield").setHidden(imageData.getColorDeconvolutionStains() == null);
map.get("excludeDAB").setHidden(imageData.getColorDeconvolutionStains() == null || !imageData.getColorDeconvolutionStains().isH_DAB());
// map.get("makeMeasurements").setHidden(!imageData.isBrightfield());
return params;
}
@Override
public String getName() {
return "Watershed cell detection";
}
@Override
public String getLastResultsDescription() {
return detector == null ? "" : detector.getLastResultsDescription();
}
static class WatershedCellDetector {
private boolean refineBoundary = true; // TODO: Consider making this variable accessible
private double backgroundRadius = 15;
private double maxBackground = 0.3;
private int z = 0, t = 0;
private boolean lastRunCompleted = false;
private boolean includeNuclei = true;
private double cellExpansion = 0;
private double minArea = 0;
private double maxArea = 0;
private double medianRadius = 2;
private double sigma = 2.5;
private double threshold = 0.3;
private boolean mergeAll = true;
private boolean watershedPostProcess = true; // TODO: COMBINE WITH MERGEALL OPTION
private boolean excludeDAB = false;
private boolean smoothBoundaries = false;
// private boolean limitExpansionByNucleusSize = false;
private boolean makeMeasurements = true;
private Roi roi = null;
private FloatProcessor fpDetection = null;
private Map<String, FloatProcessor> channels = new LinkedHashMap<>(); // Map of channels to measure for nuclei only, and their names
private Map<String, FloatProcessor> channelsCell = new LinkedHashMap<>(); // Map of channels to measure for cell/cytoplasm, and their names
private ImageProcessor ipToMeasure = null;
private List<PolygonRoi> rois = null;
private ByteProcessor bpLoG = null;
private List<PolygonRoi> roisNuclei = new ArrayList<>();
private List<PathObject> pathObjects = new ArrayList<>();
private PathImage<ImagePlus> pathImage = null;
public WatershedCellDetector(FloatProcessor fpDetection, Map<String, FloatProcessor> channels, Map<String, FloatProcessor> channelsCell, Roi roi, PathImage<ImagePlus> pathImage) {
this.fpDetection = fpDetection;
if (channels != null)
this.channels.putAll(channels);
if (channelsCell != null)
this.channelsCell.putAll(channelsCell);
this.roi = roi;
this.pathImage = pathImage;
Prefs.setThreads(1);
}
public static ByteProcessor limitedOpeningByReconstruction(final ImageProcessor ip, final ImageProcessor ipBackground, final double radius, final double maxBackground) {
// Apply (initial) morphological opening
final RankFilters rf = new RankFilters();
ipBackground.setRoi(ip.getRoi());
rf.rank(ipBackground, radius, RankFilters.MIN);
// Mask out any above-threshold background pixels & their surroundings
ByteProcessor bpMask = null;
if (!Double.isNaN(maxBackground) && maxBackground > 0) {
int w = ip.getWidth();
int h = ip.getHeight();
for (int i = 0; i < w * h; i++) {
if (ipBackground.getf(i) > maxBackground) {
if (bpMask == null)
bpMask = new ByteProcessor(w, h);
bpMask.setf(i, 1f);
}
}
// Apply mask if required
if (bpMask != null) {
rf.rank(bpMask, radius*2, RankFilters.MAX);
for (int i = 0; i < w * h; i++) {
if (bpMask.getf(i) != 0f) {
ipBackground.setf(i, Float.NEGATIVE_INFINITY);
}
}
}
}
// Apply the morphological reconstruction
MorphologicalReconstruction.morphologicalReconstruction(ipBackground, ip);
return bpMask;
}
private void doDetection(boolean regenerateROIs) {
int width = fpDetection.getWidth();
int height = fpDetection.getHeight();
// Prefs.setThreads(1);
lastRunCompleted = false;
pathObjects.clear();
ByteProcessor bp = null;
ByteProcessor bpBackgroundMask = null;
fpDetection.setRoi(roi);
if (regenerateROIs) {
rois = null;
bpLoG = null;
// Use Laplacian of Gaussian filtering followed by watershed transform to determine possible nucleus segments
// Result will be a dramatic over-segmentation...
FloatProcessor fpLoG = (FloatProcessor)fpDetection.duplicate();
// Start off with a median filter to reduce texture, if necessary
RankFilters rf = new RankFilters();
if (medianRadius > 0)
rf.rank(fpLoG, medianRadius, RankFilters.MEDIAN);
//--------NEW--------
if (excludeDAB && channels.containsKey("Hematoxylin OD") && channels.containsKey("DAB OD")) {
// If we are avoiding DAB, set pixels away from potential nuclei to zero
FloatProcessor fpDAB = channels.get("DAB OD");
fpDAB.setRoi(roi);
ByteProcessor bpH = SimpleThresholding.greaterThanOrEqual(channels.get("Hematoxylin OD"), fpDAB);
bpH.multiply(1.0/255.0);
rf.rank(bpH, 2.5, RankFilters.MEDIAN);
rf.rank(bpH, 2.5, RankFilters.MAX);
fpLoG.copyBits(bpH, 0, 0, Blitter.MULTIPLY);
}
//--------END_NEW--------
// Subtract background first, if needed
if (backgroundRadius > 0) {
ImageProcessor ipBackground = fpLoG.duplicate();
bpBackgroundMask = limitedOpeningByReconstruction(fpLoG, ipBackground, backgroundRadius, maxBackground);
fpLoG.copyBits(ipBackground, 0, 0, Blitter.SUBTRACT);
ipToMeasure = fpLoG.duplicate();
} else {
ipToMeasure = fpDetection;
}
// Apply (approximation of) Laplacian of Gaussian filter
fpLoG.blurGaussian(sigma);
fpLoG.convolve(new float[]{0, -1, 0, -1, 4, -1, 0, -1, 0}, 3, 3);
// Threshold the main LoG image
bpLoG = SimpleThresholding.thresholdAbove(fpLoG, 0f);
// Need to set the threshold very slightly above zero for ImageJ
// TODO: DECIDE ON USING MY WATERSHED OR IMAGEJ'S....
fpLoG.setRoi(roi);
ImageProcessor ipTemp = RegionalExtrema.findRegionalMaxima(fpLoG, 0.001f, false);
ImageProcessor ipLabels = ROILabeling.labelImage(ipTemp, 0, false);
Watershed.doWatershed(fpLoG, ipLabels, 0, false);
ipLabels.setThreshold(0.5, Double.POSITIVE_INFINITY, ImageProcessor.NO_LUT_UPDATE);
// TODO: Consider 4/8 connectivity for watershed nucleus ROIs
rois = ROILabeling.getFilledPolygonROIs(ipLabels, Wand.FOUR_CONNECTED);
if (Thread.currentThread().isInterrupted())
return;
}
if (bp == null)
bp = new ByteProcessor(width, height);
// // TODO: Consider application of an automated threshold
// if (threshold < 0) {
// ipToMeasure.resetRoi();
// ImageStatistics stats = ipToMeasure.getStatistics();
// threshold = stats.mean;// + stats.stdDev;
// logger.info("Mean threshold set: " + threshold);
// }
bp.setValue(255);
for (Roi r : rois) {
// Perform mean intensity check - skip if below threshold
ipToMeasure.setRoi(r);
double mean = ipToMeasure.getStatistics().mean;
if (mean <= threshold) {
continue;
}
// Perform background intensity check, if required
if (bpBackgroundMask != null) {
bpBackgroundMask.setRoi(r);
if (bpBackgroundMask.getStatistics().mean > 0)
continue;
}
// Fill the ROI to keep it
bp.fill(r);
}
if (Thread.currentThread().isInterrupted())
return;
// Create a new, updated binary image with the potential nucleus regions & (optionally) merge these
bp.setThreshold(127, Double.POSITIVE_INFINITY, ImageProcessor.NO_LUT_UPDATE);
if (mergeAll) {
bp.filter(ImageProcessor.MAX);
bp.copyBits(bpLoG, 0, 0, Blitter.AND);
if (watershedPostProcess) {
// TODO: ARRANGE A MORE EFFICIENT FILL HOLES
List<PolygonRoi> rois2 = ROILabeling.getFilledPolygonROIs(bp, Wand.FOUR_CONNECTED);
bp.setValue(255);
for (Roi r : rois2)
bp.fill(r);
new EDM().toWatershed(bp);
}
}
// TODO: Look at the better boundary clearing implemented in Fast_nucleus_counts
if (roi != null)
ROILabeling.clearOutside(bp, roi);
// Locate nucleus ROIs
bp.setThreshold(127, Double.POSITIVE_INFINITY, ImageProcessor.NO_LUT_UPDATE);
if (IJ.debugMode) {
IJTools.quickShowImage("Binary", bp.duplicate());
}
//----------------------------
// MINOR BOUNDARY REFINEMENT
// The idea is that Gaussian smoothing tends to cause the boundaries of 'thin' nuclei to be overestimated;
// this uses a smaller filter to correct instances where the boundary has moved by just one pixel
if (refineBoundary && sigma > 1.5) {
FloatProcessor fpBoundaryCleanup = (FloatProcessor)fpDetection.duplicate();
fpBoundaryCleanup.blurGaussian(1);
fpBoundaryCleanup.convolve(new float[]{0, -1, 0, -1, 4, -1, 0, -1, 0}, 3, 3);
ByteProcessor bp2 = SimpleThresholding.thresholdAbove(fpBoundaryCleanup, 0f);
bp2.copyBits(bp, 0, 0, Blitter.MIN); // Remove everything not detected in bp
bp.filter(ByteProcessor.MIN);
bp.copyBits(bp2, 0, 0, Blitter.MAX);
regenerateROIs = true;
}
//----------------------------
roisNuclei = ROILabeling.getFilledPolygonROIs(bp, Wand.FOUR_CONNECTED);
if (Thread.currentThread().isInterrupted())
return;
// Remove nuclei with areas outside the permitted range - updating the binary image as we go
if (minArea > 0 || maxArea > 0) {
bp.setValue(0);
Iterator<PolygonRoi> iter = roisNuclei.iterator();
while (iter.hasNext()) {
Roi roiTemp = iter.next();
ipToMeasure.setRoi(roiTemp);
ImageStatistics stats = ImageStatistics.getStatistics(ipToMeasure, Measurements.AREA | Measurements.MEAN, null);
double area = stats.pixelCount;
if ((stats.mean < threshold) || (minArea > 0 && area < minArea) || (maxArea > 0 && area > maxArea)) {
iter.remove();
bp.fill(roiTemp);
}
}
ipToMeasure.resetRoi();
}
// Label nuclei
ShortProcessor ipLabels = new ShortProcessor(width, height);
ROILabeling.labelROIs(ipLabels, roisNuclei);
// Measure nuclei for all required channels
Map<String, List<RunningStatistics>> statsMap = new LinkedHashMap<>();
if (makeMeasurements) {
SimpleImage imgLabels = new PixelImageIJ(ipLabels);
for (String key : channels.keySet()) {
List<RunningStatistics> statsList = StatisticsHelper.createRunningStatisticsList(roisNuclei.size());
StatisticsHelper.computeRunningStatistics(new PixelImageIJ(channels.get(key)), imgLabels, statsList);
statsMap.put(key, statsList);
}
}
if (Thread.currentThread().isInterrupted())
return;
// Create nucleus objects
// TODO: Set the measurement capacity to improve efficiency
List<PathObject> nucleiObjects = new ArrayList<>();
Calibration cal = pathImage.getImage().getCalibration();
for (int i = 0; i < roisNuclei.size(); i++) {
PolygonRoi rOrig = roisNuclei.get(i);
PolygonRoi r = rOrig;
if (smoothBoundaries)
r = new PolygonRoi(rOrig.getInterpolatedPolygon(Math.min(2.5, rOrig.getNCoordinates()*0.1), true), Roi.POLYGON);
PolygonROI pathROI = ROIConverterIJ.convertToPolygonROI(r, cal, pathImage.getDownsampleFactor(), 0, z, t);
if (smoothBoundaries) {
pathROI = ShapeSimplifier.simplifyPolygon(pathROI, pathImage.getDownsampleFactor()/4.0);
}
// Create a new shared measurement list
MeasurementList measurementList = MeasurementListFactory.createMeasurementList(makeMeasurements ? 30 : 0, MeasurementList.TYPE.FLOAT);
if (makeMeasurements) {
ObjectMeasurements.addShapeStatistics(measurementList, r, fpDetection, cal, "Nucleus: ");
for (String key : channels.keySet()) {
List<RunningStatistics> statsList = statsMap.get(key);
RunningStatistics stats = statsList.get(i);
measurementList.addMeasurement("Nucleus: " + key + " mean", stats.getMean());
measurementList.addMeasurement("Nucleus: " + key + " sum", stats.getSum());
measurementList.addMeasurement("Nucleus: " + key + " std dev", stats.getStdDev());
measurementList.addMeasurement("Nucleus: " + key + " max", stats.getMax());
measurementList.addMeasurement("Nucleus: " + key + " min", stats.getMin());
measurementList.addMeasurement("Nucleus: " + key + " range", stats.getRange());
}
}
// TODO: It would be more efficient to measure the hematoxylin intensities along with the shapes
PathObject pathObject = new PathDetectionObject(pathROI, null, measurementList);
nucleiObjects.add(pathObject);
}
if (Thread.currentThread().isInterrupted())
return;
List<Roi> roisCellsList = null;
// Optionally expand the nuclei to become cells
if (cellExpansion > 0) {
FloatProcessor fpEDM = new EDM().makeFloatEDM(bp, (byte)255, false);
fpEDM.multiply(-1);
double cellExpansionThreshold = -cellExpansion;
// Create cell ROIs
ImageProcessor ipLabelsCells = ipLabels.duplicate();
Watershed.doWatershed(fpEDM, ipLabelsCells, cellExpansionThreshold, false);
PolygonRoi[] roisCells = ROILabeling.labelsToFilledROIs(ipLabelsCells, roisNuclei.size());
// Compute cell DAB stats
Map<String, List<RunningStatistics>> statsMapCell = new LinkedHashMap<>();
if (makeMeasurements) {
for (String key : channelsCell.keySet()) {
List<RunningStatistics> statsList = StatisticsHelper.createRunningStatisticsList(roisNuclei.size());
StatisticsHelper.computeRunningStatistics(new PixelImageIJ(channelsCell.get(key)), new PixelImageIJ(ipLabelsCells), statsList);
statsMapCell.put(key, statsList);
}
}
// Create labelled image for cytoplasm, i.e. remove all nucleus pixels
// TODO: Make a buffer zone between nucleus and cytoplasm!
for (int i = 0; i < ipLabels.getWidth() * ipLabels.getHeight(); i++) {
if (ipLabels.getf(i) != 0)
ipLabelsCells.setf(i, 0f);
}
// Compute cytoplasm stats
Map<String, List<RunningStatistics>> statsMapCytoplasm = new LinkedHashMap<>();
if (makeMeasurements) {
for (String key : channelsCell.keySet()) {
List<RunningStatistics> statsList = StatisticsHelper.createRunningStatisticsList(roisNuclei.size());
StatisticsHelper.computeRunningStatistics(new PixelImageIJ(channelsCell.get(key)), new PixelImageIJ(ipLabelsCells), statsList);
statsMapCytoplasm.put(key, statsList);
}
}
// Create cell objects
roisCellsList = new ArrayList<>(roisCells.length); // In case we need texture measurements, store all cell ROIs
for (int i = 0; i < roisCells.length; i++) {
PolygonRoi r = roisCells[i];
if (r == null)
continue;
if (smoothBoundaries)
r = new PolygonRoi(r.getInterpolatedPolygon(Math.min(2.5, r.getNCoordinates()*0.1), false), Roi.POLYGON); // TODO: Check this smoothing - it can be troublesome, causing nuclei to be outside cells
// r = smoothPolygonRoi(r);
PolygonROI pathROI = ROIConverterIJ.convertToPolygonROI(r, pathImage.getImage().getCalibration(), pathImage.getDownsampleFactor(), 0, z, t);
if (smoothBoundaries)
pathROI = ShapeSimplifier.simplifyPolygon(pathROI, pathImage.getDownsampleFactor()/4.0);
MeasurementList measurementList = null;
PathObject nucleus = null;
if (includeNuclei) {
// Use the nucleus' measurement list
nucleus = nucleiObjects.get(i);
measurementList = nucleus.getMeasurementList();
} else {
// Create a new measurement list
measurementList = MeasurementListFactory.createMeasurementList(makeMeasurements ? 12 : 0, MeasurementList.TYPE.GENERAL);
}
// Add cell shape measurements
if (makeMeasurements) {
ObjectMeasurements.addShapeStatistics(measurementList, r, fpDetection, pathImage.getImage().getCalibration(), "Cell: ");
// ObjectMeasurements.computeShapeStatistics(pathObject, pathImage, fpH, pathImage.getImage().getCalibration());
// Add cell measurements
for (String key : channelsCell.keySet()) {
if (statsMapCell.containsKey(key)) {
RunningStatistics stats = statsMapCell.get(key).get(i);
measurementList.addMeasurement("Cell: " + key + " mean", stats.getMean());
measurementList.addMeasurement("Cell: " + key + " std dev", stats.getStdDev());
measurementList.addMeasurement("Cell: " + key + " max", stats.getMax());
measurementList.addMeasurement("Cell: " + key + " min", stats.getMin());
// pathObject.addMeasurement("Cytoplasm: " + key + " range", stats.getRange());
}
}
// Add cytoplasm measurements
for (String key : channelsCell.keySet()) {
if (statsMapCytoplasm.containsKey(key)) {
RunningStatistics stats = statsMapCytoplasm.get(key).get(i);
measurementList.addMeasurement("Cytoplasm: " + key + " mean", stats.getMean());
measurementList.addMeasurement("Cytoplasm: " + key + " std dev", stats.getStdDev());
measurementList.addMeasurement("Cytoplasm: " + key + " max", stats.getMax());
measurementList.addMeasurement("Cytoplasm: " + key + " min", stats.getMin());
// pathObject.addMeasurement("Cytoplasm: " + key + " range", stats.getRange());
}
}
// Add nucleus area ratio, if available
if (nucleus != null && nucleus.getROI() instanceof PathArea) {
double nucleusArea = ((PathArea)nucleus.getROI()).getArea();
double cellArea = pathROI.getArea();
measurementList.addMeasurement("Nucleus/Cell area ratio", Math.min(nucleusArea / cellArea, 1.0));
// measurementList.addMeasurement("Nucleus/Cell expansion", cellArea - nucleusArea);
}
}
// Create & store the cell object
PathObject pathObject = new PathCellObject(pathROI, nucleus == null ? null : nucleus.getROI(), null, measurementList);
pathObjects.add(pathObject);
roisCellsList.add(r);
}
} else {
pathObjects.addAll(nucleiObjects);
}
// Close the measurement lists
for (PathObject pathObject : pathObjects)
pathObject.getMeasurementList().closeList();
lastRunCompleted = true;
}
private static PolygonRoi smoothPolygonRoi(PolygonRoi r) {
FloatPolygon poly = r.getFloatPolygon();
FloatPolygon poly2 = new FloatPolygon();
int nPoints = poly.npoints;
for (int i = 0; i < nPoints; i += 2) {
int iMinus = (i + nPoints - 1) % nPoints;
int iPlus = (i + 1) % nPoints;
poly2.addPoint((poly.xpoints[iMinus] + poly.xpoints[iPlus] + poly.xpoints[i])/3,
(poly.ypoints[iMinus] + poly.ypoints[iPlus] + poly.ypoints[i])/3);
}
// return new PolygonRoi(poly2, r.getType());
return new PolygonRoi(poly2, Roi.POLYGON);
}
public List<PathObject> getPathObjects() {
return pathObjects;
}
// public void runDetection(double backgroundRadius, double maxBackground, double medianRadius, double sigma, double threshold, double minArea, double maxArea, boolean mergeAll, boolean watershedPostProcess, boolean excludeDAB, double cellExpansion, boolean limitExpansionByNucleusSize, boolean smoothBoundaries, boolean includeNuclei, boolean makeMeasurements) {
public void runDetection(double backgroundRadius, double maxBackground, double medianRadius, double sigma, double threshold, double minArea, double maxArea, boolean mergeAll, boolean watershedPostProcess, boolean excludeDAB, double cellExpansion, boolean smoothBoundaries, boolean includeNuclei, boolean makeMeasurements, int z, int t) {
boolean updateNucleusROIs = rois == null || bpLoG == null;
updateNucleusROIs = updateNucleusROIs ? updateNucleusROIs : this.medianRadius != medianRadius;
this.medianRadius = medianRadius;
updateNucleusROIs = updateNucleusROIs ? updateNucleusROIs : this.t != t || this.z != z;
this.z = z;
this.t = t;
updateNucleusROIs = updateNucleusROIs ? updateNucleusROIs : this.backgroundRadius != backgroundRadius;
this.backgroundRadius = backgroundRadius;
updateNucleusROIs = updateNucleusROIs ? updateNucleusROIs : this.sigma != sigma;
this.sigma = sigma;
updateNucleusROIs = updateNucleusROIs ? updateNucleusROIs : this.excludeDAB != excludeDAB;
this.excludeDAB = excludeDAB;
boolean updateAnything = updateNucleusROIs || !lastRunCompleted;
updateAnything = updateAnything ? updateAnything : this.minArea != minArea;
this.minArea = minArea;
updateAnything = updateAnything ? updateAnything : this.maxArea != maxArea;
this.maxArea = maxArea;
updateAnything = updateAnything ? updateAnything : this.maxBackground != maxBackground;
this.maxBackground = maxBackground;
updateAnything = updateAnything ? updateAnything : this.threshold != threshold;
this.threshold = threshold;
updateAnything = updateAnything ? updateAnything : this.mergeAll != mergeAll;
this.mergeAll = mergeAll;
updateAnything = updateAnything ? updateAnything : this.watershedPostProcess != watershedPostProcess;
this.watershedPostProcess = watershedPostProcess;
updateAnything = updateAnything ? updateAnything : this.cellExpansion != cellExpansion;
this.cellExpansion = cellExpansion;
updateAnything = updateAnything ? updateAnything : this.smoothBoundaries != smoothBoundaries;
this.smoothBoundaries = smoothBoundaries;
updateAnything = updateAnything ? updateAnything : this.includeNuclei != includeNuclei;
this.includeNuclei = includeNuclei;
updateAnything = updateAnything ? updateAnything : this.makeMeasurements != makeMeasurements;
this.makeMeasurements = makeMeasurements;
// updateAnything = updateAnything ? updateAnything : this.limitExpansionByNucleusSize != limitExpansionByNucleusSize;
// this.limitExpansionByNucleusSize = limitExpansionByNucleusSize;
// if (!updateAnything)
// return;
doDetection(updateNucleusROIs);
}
}
@Override
public String getDescription() {
return "Default cell detection algorithm for brightfield images with nuclear or cytoplasmic staining";
}
@Override
protected double getPreferredPixelSizeMicrons(ImageData<BufferedImage> imageData, ParameterList params) {
return CellDetector.getPreferredPixelSizeMicrons(imageData, params);
}
@Override
protected ObjectDetector<BufferedImage> createDetector(ImageData<BufferedImage> imageData, ParameterList params) {
return new CellDetector();
}
@Override
protected int getTileOverlap(ImageData<BufferedImage> imageData, ParameterList params) {
double pxSize = imageData.getServer().getAveragedPixelSizeMicrons();
if (Double.isNaN(pxSize))
return params.getDoubleParameterValue("cellExpansion") > 0 ? 25 : 10;
double cellExpansion = params.getDoubleParameterValue("cellExpansionMicrons") / pxSize;
int overlap = cellExpansion > 0 ? (int)(cellExpansion + 10) : 10;
// System.out.println("Tile overlap: " + overlap + " pixels");
return overlap;
}
}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment