Segment structures in the image, if yes a pixel classifier has to be defined
pixelClassifier
A SVM which classifies RGB color values in foreground and background.
greyscaleImage
Channel of the RGB image, to use for thresholding, if 0 use a joined greyscale image.
penClassifier
Classifier to exclude low quality images(will be part of next release)
referenceHist
Color histogram of a reference image, that can be used to estimate the quality of the recent image (will be part of next release)
Details
The image is converted to greyscale and thresholded.
Clutter is deleted using morphological operations. Clustered objects are separated using watershed algorithm.
Segmented Cell nuclei, which exceed the maximum size are thresholded and segmented again. Cell nuclei which fall below the minimum size are
deleted. Dark regions which exceed the parameter failureRegion are considered as artefacts and deleted.
If the parameters are not defined, the operations will not be executed.
Features are generated for every segmented object.