Watershed Segmentation ( IAAdvSegment )

Parameters:

Name	Short Name	Type	Description
Threshold value calculation method	type	string	How is the threshold level calculated: "Pixel value", "Percentile by frames", "Percentile by series", "Percentile by reference", "Otsu by frames", "Otsu by series", "Otsu by reference"
Way	way	string	"Above" for bright objects or "Below" for dark objects.
Value	value	real	Pixel value, percentile, or factor times Otsu optimal threshold value (typically 1)
Threshold from local max/min	proctype	string	"Bound uniformly" and "Bound locally" enables local maximum or minimum search. Set None for global thresholding.
Determine boundaries at	value2	real	For "Bound locally", the boundary of each object is determined at this % of the maximal intensity of the object relative to its neighborhood. For "Bound uniformly" this is a pixel intensity value.
Method	method	string	The method of separating objects from each other: "Mark Seed", "Mark Shape", "Watershed(que)", "Watershed(distance)", "Gradient".
Connectivity	connectivity	string	How the objects are propagated from the seeds: "Inf (3x3 rectangular)", "L4 (3x3 cross)", "L2", "Fast Marching"
Separate segments	separate	boolean	If yes segments will not touch each other.
Weld segments into round objects	weld	boolean	Weld touching segments if they form a rounder object together. Use this to avoid objects fragmenting into multiple segments.
Evaluate segment data	evaluate	boolean	If yes measures segments and applies classifiers, if no only marks segments with different values.
Data/Text Window Output	textoutput	boolean	Whether to show segmentation results data in a text or Excel Data window. If "No" the segment data are still available for plotting.
Keep original mask	keeporigmask	boolean	Originally masked areas are kept masked if Yes. Otherwise the mask will take zero, ceiling or bottom value.
Discard segments at edges	noedges	boolean	Any segment that has at least 10% of its boundary at the edge of the image will be discarded.

Description:

Watershed Segmentation (formerly Advanced Segmentation). Segments grayscale images using the Watershed algorithm. The threshold value calculation methods are identical to the ones used for the Threshold function.
Way: Above segments bright details, Below segments dark details as objects
Segmentation is performed from local maxima or minima for "Above" or "Below" setting, respectively at the "Way" parameter.
Threshold from local max/min: "None" calculates simple threshold according to the threshold value, "Bound uniformly" calculates segment boundary with morphological reconstruction using a cap value modified by the "Determine boundaries at" value. Bound locally will bound local maxima at their given percent compared to the minimum or maximum of the image. This is performed by flood filling from the local maximum or minimum, respecting the boundaries of the Watershed segments.
Determine boundaries at: when "Bound uniformly" is selected above, given in the same units and additive to the threshold value, and sets how far from the local maximum will the boundary of the object drawn. When "Bound locally" is selected, it is a value in percents of local maximum or minimum.
Method: "Mark Seed" shows local maxima or minima by each as a single pixel. "Mark Shape" shows the boundaries of the objects, undivided by the watershed algorithm. Watershed divides bounded segments propagating from local maxima or minima.
The number of segments is limited only by the available memory.
Note: the segmentation algorithm has integer core, and does not work properly if the image consist of small, fractional values. Use Scalar Arithmetic to set a proper range, where pixel values are >>1.
The "Keep original mask" is not applicable here; masked areas are ignored during threshold calculation, but substituted by zeros in the segmented images.