How to segment out and label lymphatic vessels on histology image?

histology
java
python
segmentation
Tags: #<Tag:0x00007fb87cc48448> #<Tag:0x00007fb87cc481c8> #<Tag:0x00007fb87d79bf98> #<Tag:0x00007fb87d79bd18>

#1

How to better segment out and label brown-colored lymphatic vessels on a histology image?

Need to choose a proper conversion, segmentation, labelling methods and respective image processing libraries.

Example is an image at http://www.jocmr.org/index.php/JOCMR/article/viewFile/749/386/4153

  I have tried Java Open CV code:
         package javaopencvapp;
        import org.opencv.core.Core;
        import org.opencv.core.Mat;
        import org.opencv.core.CvType; 
        import org.opencv.core.Scalar;
        import javaopencvapp.ImageViewer;
       import org.opencv.core.Core;
       import org.opencv.core.Mat;
       import org.opencv.imgcodecs.Imgcodecs;
       import org.opencv.imgproc.Imgproc;

  public class JavaOpenCVapp {

    static{ System.loadLibrary(Core.NATIVE_LIBRARY_NAME); }
    public static void main(String[] args) throws Exception {
	String filePath = "src/images/imth2.tif";
	Mat newImage = Imgcodecs.imread(filePath);
            
            Imgproc.threshold(newImage, newImage, 172, 179.0, Imgproc.THRESH_BINARY);
           //* Imgproc.threshold(newImage, newImage, 0, 125, Imgproc.THRESH_BINARY + Imgproc.THRESH_OTSU);
           //*  Imgproc.adaptiveThreshold(newImage, newImage, 255, Imgproc.ADAPTIVE_THRESH_GAUSSIAN_C, Imgproc.THRESH_BINARY, 11, 2); 
      //* Imgproc.adaptiveThreshold(newImage,newImage, 255, Imgproc.ADAPTIVE_THRESH_MEAN_C,Imgproc. THRESH_BINARY, 15, 40);
           
	if(newImage.dataAddr()==0){
		System.out.println("Couldn't open file " + filePath);
	}else{
		ImageViewer imageViewer;
                    imageViewer = new ImageViewer();
		imageViewer.show(newImage, "Loaded image");
	}

}

}

thank you


#2

Sorry to come so late to your post. Did you find a solution? If not, you might want to give the Trainable Weka Segmentation a try.


#3

Did you ever find a solution to this?

Thanks,
Samir