Module: Merge ()

Description:

This module works on any 3-dimensional field on rectilinear coordinates and merges the input data by interpolation.

Merging

Proceed as follows: Attach a Merge module to an input field. For additional inputs, right-click on the white square on the left side of the Merge module's icon, select an input item, and drag a connection to a data icon. Input fields can have an arbitrary transformation in which case the result field will have an axis aligned that spans the bounding boxes of the input fields. The voxel size of the result field is determined by the first input. At locations within the resulting volume where none of the input fields are defined voxels are initialized with zeros.

For image data, you have the choice among various interpolation methods:

• Nearest Neighbor chooses for every new voxel the average value of the input voxels nearest to it.
• Standard linearily interpolates between the surrounding voxels.
• Lanczos is the slowest but most accurate method and approximates a low pass filter. If you have time and want to get the best result, use this one.

For label images, this port is hidden and always Nearest Neighbor is used.

Choose an additional option and press the Apply button.

Merging in Progress

While merging, the progress bar at the bottom of the Properties Area indicates the percentage of merging that is done. You may cancel the merging any time by pressing the stop button on the right of the progress bar.

Connections:

Data [required]
The first data field to be merged.

Lattice* [optional]
Additional input data.

Ports:

Interpolation

The interpolation method that will be used. There are three methods available:

• Nearest Neighbor interpolation:
  Actually this method does not interpolate field values but returns the data values stored at the grid point nearest to the position where the field is to be evaluated. Only grid points belonging to the grid cell the position vector lies in are considered. Nearest-neighbor lookup is particularly useful for labels, where standard interpolation makes no sense.
• Standard interpolation:
  This method interpolates a field using its standard interpolation method. Usually, for fields with regular coordinates this is trilinear interpolation.
• Lanczos interpolation:
  This method does high-quality interpolation using the Lanczos3 kernel. This kernel approximates a lowpass sinc filter. Currently, the method is only available for regular grids with uniform coordinates. Note, that this type of interpolation is much more expensive than standard trilinear filtering.

Options

Option blend controls the merging process. If this toggle is disabled, a voxel value of the first data field is used if such exists at a given location. Otherwise, a result value will be calculated from the additional input data. If this toggle is enabled, result values are always interpolated. If all input data are label images, this option is automatically disabled.

Use the second option use existing result to overwrite an existing result. If use existing result is not set, a new result object is generated.

Padding Value

Padding value is used for monochrome images to define the value used to fill in undefined pixels.

Padding Color

Padding value is used for RGBA images to define the value used to fill in undefined pixels.