ConstrainedConnectionCost

MLModule

genre	Morphology
author	MeVis Medical Solutions AG
package	MeVisLab/Standard
dll	MLRate
definition	MLRate.def
see also	Morphology, BitMorphology, CloseGap
keywords	morphology, distance, fill

Purpose

The module ConstrainedConnectionCost calculates 2D connection cost of image voxels with respect to a given set of reference voxels.

Depth constraints of a marker image are taken into account. Connection costs are computed independently for each x-y slice of the input data.

Given the path cost as the maximum image value over all voxels belonging to a given path, the connection cost of two voxels is defined as the minimum path cost concerning all possible connections between them.

Hence, the connection cost between two points x and y corresponds to the minimum barrier that must be crossed in order to pass from x to y.

Starting at the set of reference voxels, the algorithm results in filling all the local valleys present in the image.

Usage

Computes an image of minimum values (CCC values) that any path from the boundary needs to overcome to reach a voxel in a slice.

The result is displayed directly or used to show the difference between this value and the original or to mask out parts of the image.

Details

The reference set for the connection cost consists of all voxels v where input1(v) is less than input0(v). For the other voxels, input1 provides a local upper constraint for the calculated connection cost.

If no reference image is used, connection costs are calculated with respect to the border of image input0 (i.e., for filling holes in binary masks).

The module works globally in the x-y plane; hence, the page extent is set appropriately.

Windows

Default Panel

Input Fields

input0

name: input0, type: Image

Primary input: image data. Common gray scale images or binary images are typical use cases

input1

name: input1, type: Image

Marker image for use with the use marker image option, an alternative to computing connection costs with respect to the image boundary.

Output Fields

output0

name: output0, type: Image

The resulting image

Parameter Fields

Field Index

Compute Diff Image: Bool

Compute Masked Image: Bool

Connection: Enum

Fill Value: Double

Marker Status (input1): String

Mask Orig Image: Bool

Min CCCValue: Double

Use marker (input1): Bool

Visible Fields

Use marker (input1)

name: useMarker, type: Bool, default: FALSE, deprecated name: UseMarker

If checked, input1 is used for constraining the connection cost.

Otherwise, connection costs are calculated with respect to the border of image input0 (i.e., for filling gaps in binary masks).

Compute Diff Image

name: computeDiffImage, type: Bool, default: FALSE

If checked, the difference image is computed.

Compute Masked Image

name: computeMaskedImage, type: Bool, default: FALSE

If checked, this option will mask the output so that only parts where the CCC exceeds Min CCCValue are present in the output image.

Voxels with a CCC below Min CCCValue will be assigned the specified :field:fillValue; the rest of the output depends on further settings (i.e., one will receive either the CCC values, the original values, or the difference between the CCC and original values).

Min CCCValue

name: minCCCValue, type: Double, default: 1

Sets the threshold of CCC used in masking.

Fill Value

name: fillValue, type: Double, default: 0

Sets the fill value for the background region in the masking mode.

Mask Orig Image

name: maskOrigImage, type: Bool, default: FALSE

If checked, the masked original image will be displayed in the output for voxels with a CCC above Min CCCValue, instead of showing CCC values.

Only used if Compute Masked Image is active.

Marker Status (input1)

name: markerStatus, type: String, persistent: no, deprecated name: MarkerStatus

Shows information about the marker input image.

Connection

name: connectionMode, type: Enum, default: ConnectXY, deprecated name: Connection

Defines the path along which the connection cost are evaluated.

For standard purposes, all paths in the x-y plane are feasible; however, using only paths parallel to the x or y-axis may be useful in specific cases.

Values:

Title	Name	Deprecated Name	Description
xy-plane-8nb	ConnectXY	xy-plane	Connections over all eight 2D neighbor voxels are considered.
xy-plane-4nb	ConnectXY4NB		Connections over only the closest four 2D neighbor voxels are considered.
x-axis	ConnectX	x-axis	Connections over only the two neighbor voxels in 1D rows are considered.
y-axis	ConnectY	y-axis	Connections over only the two neighbor voxels in 1D columns are considered.