Description:

This module allows to compute the per-voxel average of an arbitrary number of 3D image data sets or label fields. The data sets are loaded from disk successively so that they do not have to fit into main memory simultaneously. A transformation per data set can be taken into account. The transformation matrix needs to be stored along with each data set as a parameter.

If label fields are ''averaged'' two modes are provided which are described below.

Connections:

Data [required]
This data set needs to be connected to a template data set. This data set defines location and resolution of the resulting field. In case of label averaging, the template must be a label field.

Ports:

Mode

Four different modes are available:

• Average computes the grey value average per voxel.
• Avg+Std.Dev. computes the grey value average and the standard deviation per voxel. This requires twice as much time as the Average option.
• Label Compute the probability map for one specific label which has to be specified. The probability map contains for each voxel the probability for that material voxel to occur at that voxel. The output ranges from 0 to 100 where 100 stands for a probability of 1.
• BestLabel This computes for each voxel the label which is most likely to occur at that point. Additionally it creates a probability map which contains the probability for the specific label at each voxel.

Material

Only used in the Label mode.

Files

A list of files to be worked on separated with space. This list can filled using the Choose Files button explained below.

Transform

The name of a parameter holding the 16 parameters of a transformation. This parameter should be present in all data sets.

In order to set such a parameter you can follow this recipe: After you have aligned a data set to your template (e.g. by using the transform editor), you can type into the Amira console window:

eval <dataset> parameters setValue <parameter name> [ <dataset> getTransform ]

(Replace <dataset> with the name of the particular data set and <parameter name> with the name you have chosen for the parameter, like e.g. MyTransformation).

Alternatively you can specify a TCL procedure by using the syntax TCL:procedure-name which will be called with the name of the data set and which should return the transformation. This allows you to store transformations in an external file or data base. This is used e.g. by the ''virtual insect brain'' scripts available at http://www.amiravis.com/vib.

Choose

Clicking on this button will open a file browser which allows you to select multiple files to be worked on.

DoIt

Start the computation. Depending on the resolution and number of data sets this can take minutes to hours.