public TissueAnnotaionList groubByOnClassAndOthers(String ClassName)
{
TissueAnnotaionList newList = new TissueAnnotaionList();
foreach (TissueAnnotationClass annotaion in this.tissueList)
{
TissueAnnotationClass newAnnotaion = annotaion.Clone();
if (!newAnnotaion.getAnnotaionClass().Equals(ClassName))
{
newAnnotaion.setAnnotaionClassOther();
}
newList.add(newAnnotaion);
}
newList.sort();
return(newList);
}
private static TissueAnnotaionList GetAnnotaions(String srcSlides, String outputPics)
{
TissueAnnotaionList annotationList = new TissueAnnotaionList();
int i = 0;
//Get Anotaions from silde
foreach (var slideName in Util.GetSlideFilenames(new String[] { srcSlides }))
{
using (var slideCache = new SlideCache(slideName))
{
Parallel.ForEach(slideCache.Slide.GetAnnotations(), (annotation) =>
{
if (abort != 0)
{
if (i >= abort)
{
return;
}
i++;
}
//Annotaions Bitmap extrahieren
var contained = new List <IAnnotation>();
foreach (var candidate in slideCache.Slide.GetAnnotations())
{
if (candidate == annotation)
{
continue;
}
var rectangle = annotation.BoundingBox;
rectangle.Intersect(candidate.BoundingBox);
if (rectangle.IsEmpty)
{
continue;
}
if (annotation.BoundingBox.Contains(candidate.BoundingBox))
{
contained.Add(candidate);
}
}
using (Bitmap annotationBitmap = annotation.Extract(1, contained))
{
TissueAnnotationClass tissueAnnotation = new TissueAnnotationClass(annotation.Id, annotation.Name, slideCache.SlideName);
//Werte Berechnsen
tissueAnnotation = tissueAnnotation.ComputeFeatureValues(annotationBitmap);
//Zur Liste hinzufügren
annotationList.add(tissueAnnotation);
//FOR DEBUG Save Image
if (isSave)
{
annotationBitmap.Save(outputPics + "\\" + tissueAnnotation + ".png");
}
Console.WriteLine(tissueAnnotation + " exc:" + contained.Count);
}
});
if (abort != 0 && i >= abort)
{
return(annotationList);
}
}
}
return(annotationList);
}
public void add(TissueAnnotationClass annotaion)
{
this.tissueList.Add(annotaion);
}
public static TissueAnnotationClass ComputeFeatures(TissueAnnotationClass tissueAnnotationOld, Bitmap image)
{
using (Bitmap hImage = image.Clone() as Bitmap, eImage = image.Clone() as Bitmap)
{
TissueAnnotationClass tissueAnnotation = tissueAnnotationOld.Clone();
int imageSize = 0;
//H und E Werte aus Bild ermitteln
var gpH = new ColorDeconvolution().Get1stStain(hImage, ColorDeconvolution.KnownStain.HaematoxylinEosin);
var gpE = new ColorDeconvolution().Get2ndStain(eImage, ColorDeconvolution.KnownStain.HaematoxylinEosin);
uint[] hHistogram = new uint[256];
uint[] eHistogram = new uint[256];
foreach (var grayscalePixel in gpH.Pixels())
{
imageSize++;
hHistogram[grayscalePixel.V]++;
eHistogram[gpE.GetPixel(grayscalePixel.X, grayscalePixel.Y)]++;
}
//Segmentation aus dem Bild ermitteln
//Layers erstellen
var luminaLayer = ObjectLayerCreate.CreateLuminaLayer(image);
var coresLayer = ObjectLayerCreate.CreateCoresLayer(image);
//Features dem layer hinzufügen
coresLayer =
ObjectLayerrUtils.AddFeatureFormFactor(
ObjectLayerrUtils.AddFeatureCenterPoint(coresLayer));
luminaLayer =
ObjectLayerrUtils.AddFeatureCoresInNear(
ObjectLayerrUtils.AddFeatureFormFactor(luminaLayer), coresLayer);
//Alle Features ermitteln und in Listen schreiben --> sortiert um anschließend Mid und Avverage auszurechen
List <uint> hHistogramList = histogramToList(hHistogram);
List <uint> eHistogramList = histogramToList(eHistogram);
List <uint> coresSizeList = ObjectLayerrUtils.GetFeatureValueList(coresLayer, "area");
List <uint> luminaSizeList = ObjectLayerrUtils.GetFeatureValueList(luminaLayer, "area");
List <uint> coresFormFactorList = ObjectLayerrUtils.GetFeatureValueList(coresLayer,
"FormFactorOfContour");
List <uint> luminaFormFactorWithSize = new List <uint>();
List <uint> luminaCoresInNear = new List <uint>();
List <uint> luminaCoresInNearWithForm = new List <uint>();
foreach (var imageObject in luminaLayer.Objects)
{
var coresInNear = imageObject.Features.GetFeatureByName("coresInNear").Value;
var formFactor = imageObject.Features.GetFeatureByName("FormFactorOfContour").Value;
var area = imageObject.Features.GetFeatureByName("area").Value;
if (Math.Abs(coresInNear) != 0)
{
luminaCoresInNear.Add((uint)(area / coresInNear));
luminaCoresInNearWithForm.Add((uint)((area * formFactor) / coresInNear));
}
if (Math.Abs(formFactor) != 0)
{
luminaFormFactorWithSize.Add((uint)(area / (2 * formFactor)));
}
}
luminaCoresInNear.Sort();
luminaCoresInNearWithForm.Sort();
luminaFormFactorWithSize.Sort();
//alle Features dem Object hinzufügen
tissueAnnotation.Q25H = calcQuantil(hHistogramList, 0.25);
tissueAnnotation.MeanH = calcQuantil(hHistogramList, 0.5);
tissueAnnotation.Q75H = calcQuantil(hHistogramList, 0.75);
tissueAnnotation.Q25E = calcQuantil(eHistogramList, 0.25);
tissueAnnotation.MeanE = calcQuantil(eHistogramList, 0.5);
tissueAnnotation.Q75E = calcQuantil(eHistogramList, 0.75);
if (coresSizeList.Count == 0)
{
tissueAnnotation.CountCores = 0u;
}
else
{
tissueAnnotation.CountCores = (uint)(imageSize / coresSizeList.Count);
}
if (luminaSizeList.Count == 0)
{
tissueAnnotation.CountLumina = 0u;
}
else
{
tissueAnnotation.CountLumina = (uint)(imageSize / luminaSizeList.Count);
}
tissueAnnotation.MidCoresSize = calcMid(coresSizeList);
tissueAnnotation.MeanCoresSize = calcQuantil(coresSizeList, 0.5);
tissueAnnotation.Q25CoresSize = calcQuantil(coresSizeList, 0.25);
tissueAnnotation.Q75CoresSize = calcQuantil(coresSizeList, 0.75);
tissueAnnotation.MidLuminaSize = calcMid(luminaSizeList);
tissueAnnotation.MeanLuminaSize = calcQuantil(luminaSizeList, 0.5);
tissueAnnotation.Q25LuminaSize = calcQuantil(luminaSizeList, 0.25);
tissueAnnotation.Q75LuminaSize = calcQuantil(luminaSizeList, 0.75);
if (coresSizeList.Count == 0)
{
tissueAnnotation.DensityCores = 0;
}
else
{
tissueAnnotation.DensityCores = (uint)(imageSize /
coresSizeList.Count);
}
tissueAnnotation.MidFormFactorCores = calcMid(coresFormFactorList);
tissueAnnotation.MeanFormFactorCores = calcQuantil(coresFormFactorList, 0.5);
tissueAnnotation.Q25FormFactorCores = calcQuantil(coresFormFactorList, 0.25);
tissueAnnotation.Q75FormFactorCores = calcQuantil(coresFormFactorList, 0.75);
tissueAnnotation.MidFormFactorLuminaWithSize = calcMid(luminaFormFactorWithSize);
tissueAnnotation.MeanFormFactorLuminaWithSize = calcQuantil(luminaFormFactorWithSize, 0.5);
tissueAnnotation.Q25FormFactorLuminaWithSize = calcQuantil(luminaFormFactorWithSize, 0.25);
tissueAnnotation.Q75FormFactorLuminaWithSize = calcQuantil(luminaFormFactorWithSize, 0.75);
tissueAnnotation.MidDensityLuminaCoresInNear = calcMid(luminaCoresInNear);
tissueAnnotation.MeanDensityLuminaCoresInNear = calcQuantil(luminaCoresInNear, 0.5);
tissueAnnotation.Q25DensityLuminaCoresInNear = calcQuantil(luminaCoresInNear, 0.25);
tissueAnnotation.Q75DensityLuminaCoresInNear = calcQuantil(luminaCoresInNear, 0.75);
tissueAnnotation.MidDensityFormFactorLuminaCoresInNear = calcMid(luminaCoresInNearWithForm);
tissueAnnotation.MeanDensityFormFactorLuminaCoresInNear =
calcQuantil(luminaCoresInNearWithForm, 0.5);
tissueAnnotation.Q25DensityFormFactorLuminaCoresInNear =
calcQuantil(luminaCoresInNearWithForm, 0.25);
tissueAnnotation.Q75DensityFormFactorLuminaCoresInNear =
calcQuantil(luminaCoresInNearWithForm, 0.75);
return(tissueAnnotation);
}
}
