com.actelion.research.orbit.imageAnalysis.tasks.ExclusionMapGen Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of orbit-image-analysis Show documentation
Show all versions of orbit-image-analysis Show documentation
Orbit, a versatile image analysis software for biological image-based quantification
/*
* Orbit, a versatile image analysis software for biological image-based quantification.
* Copyright (C) 2009 - 2017 Actelion Pharmaceuticals Ltd., Gewerbestrasse 16, CH-4123 Allschwil, Switzerland.
*
* 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 classShapes = model.getExclusionModel().getClassShapes();
double scaleFactor = rFrame.bimg.getWidth() / (double) originalFrame.bimg.getWidth();
rFrame.setClassShapes(classShapes);
rFrame.setFeatureDescription(model.getExclusionModel().getFeatureDescription());
rFrame.setWindowSize(model.getExclusionModel().getFeatureDescription().getWindowSize());
exMap = new ExclusionMapGen(rFrame, model.getExclusionModel());
exMap.setScaleFactor(scaleFactor);
} catch (Exception e) {
e.printStackTrace();
}
}
return exMap;
}
/**
* Returns the recognitionFrame used for the exclusionMap.
*
* @param rf
* @param model
* @return
*/
public static RecognitionFrame getExclusionMapFrame(RawDataFile rdf, RecognitionFrame rf, OrbitModel model) {
RecognitionFrame rFrame = null;
// if MIPS are available, the lowest resolution is taken, otherwise the original frame
if (rdf != null) {
try {
rf = new RecognitionFrame(rdf, false); // load full frame (ignore lower level rf)
rf.loadAnnotationROI(rdf.getRawDataFileId(), model.getAnnotationGroup());
rf.loadImageScale(rdf.getRawDataFileId());
} catch (OrbitImageServletException e) {
e.printStackTrace();
}
}
if (rf.bimg.getMipMaps() != null) {
int numMIPS = rf.bimg.getMipMaps().length;
int selectedMip = Math.max(0, numMIPS - model.getExclusionLevel());
logger.debug("seleted MIP: " + selectedMip + " (model.exclusionLevel=" + model.getExclusionLevel() + ")");
rFrame = new RecognitionFrame(rf, selectedMip, false); // -1
} else { // clone current frame
logger.debug("no MIPS available, cloning current frame");
if (model.isApplyExclusionOnNegativeChannel() && rf.getNegativeChannel() == null) {
System.out.println("Exclusion map should be used on negative channel, but no negative channel available.");
}
if (model.isApplyExclusionOnNegativeChannel() && rf.getNegativeChannel() != null) {
rFrame = new RecognitionFrame(rf.getNegativeChannel(), false);
} else {
rFrame = new RecognitionFrame(rf, false);
}
}
return rFrame;
}
public void generateMap() throws OrbitImageServletException {
logger.trace("generateMap start");
isGenerated = false;
// classify
// in future for numThreads==1 the direct method with ClassificationTaskTiled should be used (but does not work yet, ratios are always 0...
//if (numThreads>1)
{
ClassificationWorker classificationWorker = new ClassificationWorker(recognitionFrame, model, true, null, null);
classificationWorker.setNumClassificationThreads(numThreads);
classificationWorker.setDoNormalize(doNormalize);
logger.trace("executing recognitionWorker");
classificationWorker.doWork();
} /*else {
// direct single threaded method without extra thread (used for map reduce tasks)
try {
double[] ratio = new double[model.getClassShapes().size()];
recognitionFrame.setRatio(ratio);
recognitionFrame.setClassImage(new TiledImageWriter(recognitionFrame.bimg.getWidth(), recognitionFrame.bimg.getHeight(), recognitionFrame.bimg.getTileWidth(), recognitionFrame.bimg.getTileHeight()));
ClassificationTaskTiled classificationTask = new ClassificationTaskTiled(ClassifierWrapper.makeCopy(model.getClassifier()), model.getStructure(), model.getFeatureDescription(),model.getClassShapes(), recognitionFrame.getROI(), recognitionFrame.bimg, recognitionFrame.getClassImage(), null, true);
classificationTask.setPixelFuzzyness(0d);
Long[] vals = classificationTask.call();
for (int i = 0; i < vals.length; i++) {
recognitionFrame.getRatio()[i] = vals[i];
}
} catch (Exception e) {
e.printStackTrace();
throw new OrbitImageServletException("error classifying exclusion frame");
}
} */
// a classification image is only needed if computed on different (mip-)level then original image
if (keepVisualClassificationImage) {
if (recognitionFrame.getClassImage().getImage().getWidth() * (long) recognitionFrame.getClassImage().getImage().getHeight() < 8000L * 8000L) {
classificationImage = new TiledImageWriter(PlanarImage.wrapRenderedImage(recognitionFrame.getClassImage().getImage().getAsBufferedImage()));
} else {
// bad luck...
logger.warn("classimage too big for exclusion map. Dims: " + recognitionFrame.getClassImage().getImage().getWidth() + "x" + recognitionFrame.getClassImage().getImage().getHeight());
logger.error("Exclusion map size too big. Please select a small exclusion model level.");
}
}
logger.trace("recognitionWorker finished");
recognitionFrame.initializeClassColors();
/* (1) */
/*
if (model.isPerformErodeDiliate())
{
logger.trace("start erode and diliate");
for (int i=0; i<3; i++) smap = diliate(smap); // 3
for (int i=0; i<50; i++) smap = erode(smap); //10
logger.trace("end erode and diliate");
} else {
logger.trace("erode and diliate not active");
}
*/
logger.trace("start class merging");
List exclList = new ArrayList();
List inclList = new ArrayList();
for (byte ci = 0; ci < model.getClassShapes().size(); ci++) {
if (model.getClassShapes().get(ci).getIncExcMode() == ClassShape.INCLUSION) inclList.add(ci);
else if (model.getClassShapes().get(ci).getIncExcMode() == ClassShape.EXCLUSION) exclList.add(ci);
}
if (model.getVersion() < 7) { // backward compatibility
logger.debug("model version<7 detected: using class 1+2 as exclusion and 3+4 as inclusion class");
exclList.add((byte) 0);
exclList.add((byte) 1);
inclList.add((byte) 2);
inclList.add((byte) 3);
}
recognitionFrame.changeClassLabels(exclList, (byte) 1, inclList, (byte) 0);
logger.trace("finished class merging");
// generate smap
smap = new short[recognitionFrame.getClassImage().getWidth()][recognitionFrame.getClassImage().getHeight()];
// init
Point[] tileArr = recognitionFrame.getClassImage().getImage().getTileIndices(null);
for (Point tileNum : tileArr) {
for (int x = recognitionFrame.getClassImage().getImage().tileXToX(tileNum.x); x < Math.min(recognitionFrame.getClassImage().getImage().tileXToX(tileNum.x) + recognitionFrame.getClassImage().getImage().getTileWidth(), recognitionFrame.getClassImage().getImage().getWidth()); x++) {
for (int y = recognitionFrame.getClassImage().getImage().tileYToY(tileNum.y); y < Math.min(recognitionFrame.getClassImage().getImage().tileYToY(tileNum.y) + recognitionFrame.getClassImage().getImage().getTileHeight(), recognitionFrame.getClassImage().getImage().getHeight()); y++) {
byte c = recognitionFrame.getClass(x, y);
if (c > 0) { // > 0 not background, not assigned
smap[x][y] = 1; // c0, c1 -> excluded
} else {
smap[x][y] = 0; // c2, c3
}
} //y
} // x
} // tileNum
// deactivated 17.10.2010 (activate it for lung fibrosis quantification (not mast cell count))
// TODO!!!
/*
* does not work here, maybe because of classmerging before (and getClass does not work anymore...?) Move to (1), but then smap does not exist...
* do it directly on recongitionFrame.bimg ?
*/
/*
if (model.isPerformErodeDiliate())
{
logger.trace("start erode and diliate");
for (int i=0; i<3; i++) smap = diliate(smap);
for (int i=0; i<10; i++) smap = erode(smap);
logger.trace("end erode and diliate");
} else {
logger.trace("erode and diliate not active");
}
*/
isGenerated = true;
logger.trace("generateMap end");
if (logger.isTraceEnabled()) {
int cnt0 = 0;
int cnt1 = 0;
for (int x = 0; x < smap.length; x++)
for (int y = 0; y < smap[0].length; y++) {
if (smap[x][y] == 0) cnt0++;
else cnt1++;
}
logger.trace("cnt class 0:" + cnt0 + " cnt class 1:" + cnt1);
}
}
/**
* generateMap must be called before (no check will be done).
* setScaleFactor should be set before (will not automatically be done!)
*
* @param x
* @param y
* @return if pixel x,y should be excluded or not
*/
public boolean isExcluded(int x, int y) {
if ((x * scaleFactor) >= smap.length || (y * scaleFactor) >= smap[0].length) {
//logger.trace("warning: coordinates out of range: "+nx+"/"+ny+" xy: "+x+"/"+y+" scale: "+scaleFactor);
return true;
}
//logger.trace("ex: "+nx+"/"+ny+" xy: "+x+"/"+y+" scale: "+scaleFactor+" smap: "+smap[nx][ny]);
return smap[(int) (x * scaleFactor)][(int) (y * scaleFactor)] > 0;
}
/**
* checks if a certain percent (pIncluded=0..1) is included or not
*
* @param shape
* @return
*/
public boolean isExcluded(Shape shape, double pIncluded) {
long cnt = 0;
long hit = 0;
Rectangle re = shape.getBounds();
for (int x = re.x; x < re.getMaxX(); x += 2)
for (int y = re.y; y < re.getMaxY(); y += 2) {
if (shape.contains(x, y)) {
cnt++;
if (!isExcluded(x, y)) {
hit++;
}
}
}
double pInc = (double) hit / (double) cnt;
return (pInc < pIncluded);
}
private short[][] erode(final short[][] buf) {
short cb = 0;
short cf = 1;
short[][] res = new short[buf.length][buf[0].length];
for (int x = 0; x < buf.length; x++)
for (int y = 0; y < buf[x].length; y++) {
res[x][y] = buf[x][y];
if (x > 0 && buf[x - 1][y] == cf) res[x][y] = cf;
else if (x < buf.length - 1 && buf[x + 1][y] == cf) res[x][y] = cf;
else if (y > 0 && buf[x][y - 1] == cf) res[x][y] = cf;
else if (y < buf[x].length - 1 && buf[x][y + 1] == cf) res[x][y] = cf;
}
return res;
}
private short[][] diliate(final short[][] buf) {
short cb = 1;
short cf = 0;
short[][] res = new short[buf.length][buf[0].length];
for (int x = 0; x < buf.length; x++)
for (int y = 0; y < buf[x].length; y++) {
res[x][y] = buf[x][y];
if (x > 0 && buf[x - 1][y] == cf) res[x][y] = cf;
else if (x < buf.length - 1 && buf[x + 1][y] == cf) res[x][y] = cf;
else if (y > 0 && buf[x][y - 1] == cf) res[x][y] = cf;
else if (y < buf[x].length - 1 && buf[x][y + 1] == cf) res[x][y] = cf;
}
return res;
}
public RecognitionFrame getRecognitionFrame() {
return recognitionFrame;
}
public OrbitModel getModel() {
return model;
}
public boolean isGenerated() {
return isGenerated;
}
public double getScaleFactor() {
return scaleFactor;
}
public void setScaleFactor(double scaleFactor) {
this.scaleFactor = scaleFactor;
}
public boolean useForSegmentation() {
return model.isUseExclusionForSegmentation();
}
public TiledImageWriter getClassificationImage() {
return classificationImage;
}
public void setClassificationImage(TiledImageWriter classificationImage) {
this.classificationImage = classificationImage;
}
public boolean isDoNormalize() {
return doNormalize;
}
public void setDoNormalize(boolean doNormalize) {
this.doNormalize = doNormalize;
}
public int getNumThreads() {
return numThreads;
}
public void setNumThreads(int numThreads) {
this.numThreads = numThreads;
}
public boolean isKeepVisualClassificationImage() {
return keepVisualClassificationImage;
}
public void setKeepVisualClassificationImage(boolean keepVisualClassificationImage) {
this.keepVisualClassificationImage = keepVisualClassificationImage;
}
}