Image objects are organized in one or multiple image object levels per image. An image object level serves as a layered working area of processes; here you commonly apply most algorithms.
An image object level can be created by segmentation algorithms from the underlying pixel level or from an existing image object level.
Algorithms Overview on page 155
(drop-down arrow button)
Reference Handbook
Use Thresholds for Classification with Class Descriptions on page 212
In addition, you can create an image object level by duplicating an existing image object level.
Duplicate Image Object Level on page 195
Multiple image objects levels of an image perform an image object hierarchy.
Image Object Hierarchy on page 26
Keep in mind the important difference of image object level and image layer: While an image object level serves as working area, the image layer represents the data base of of an image analysis.
1. To specify the Image Object Domain `, select an image object level or another general domain in the drop-down list box. The following general domains are available:
Domain Usage
no image object For algorithms that do not use pixels or image objects (for example, Execute child process or Update variable)
pixel level Select to apply the algorithm on the pixel level (for example, segmentation algorithms).
Select to apply the algorithm on the currently active image object level.
Click the Parameters ` button to select another image object level in the Select Level dialog box.
Figure 109: Select Level dialog box.
Use Hierarchical Image Object Levels on page 193
If the chosen algorithm creates a new image object level, you may want to avoid the automatic naming of new image object levels. In that case you can define a new name beforehand. It is assigned to the image object level during process execution. Right-click in the Select Level dialog box and select Create New Level Name. In the opening dialog box you can enter the name of the new image object level.
Edit an Image Object Level or Level Variable on page 196
current image object
Select to apply the algorithm to the current internal selected image object of the parent process.
neighbor obj.
(dist=0)
Apply the algorithm to all neighbors of the current internal selected image object of the parent process. The size of the neighborhood is defined by the distance value, the spatial distance. Default value 0 adds all direct neighboring image objects. Press the Parameters ` button to set the distance in the Number Input dialog box. The unit is pixel, regardless to the unit defined when creating the project.
Figure 110: Number Input dialog box.
Spatial Distance on page 172
sub- objects (down=1)
Apply the algorithm to all subobjects of the current internal selected image object of the parent process. The number of levels down the image objects level hierarchy is defined by the down value. Press the Parameters ` button to set the steps down in the Number Input dialog box.
Level Distance on page 172
super objects (up=1)
Apply the algorithm to the superobject of the current internal selected image object of the parent process. The number of levels up the image objects level hierarchy is defined by the up value. Press the
Parameters ` button to set the steps up in the Number Input dialog box.
Level Distance on page 172
2. To specify the Image Object Domain, you can select one or more classes for further narrowing down the image objects of interest.
2.1 Click the broad All Objects a button (alternatively displaying the names of selected classes) to open the Edit Classification Filter dialog box.
Figure 111: Edit Classification Filter dialog box.
2.2 Select the classes to which the algorithm will apply:
• unclassified image objects
• one class
• multiple classes. To select, click the List button. Hold the Ctrl key while selecting multiple classes with the cursor.
2.3 The selected classes operate as a filter to the image object domain. After confirmation with OK the selected classes are displayed on the broad button a. Default setting is to use all image objects of the domain that is selected in the drop-down list.
3. To specify the Image Object Domain, you further can define a condition based on a feature. This allows you to limit the execution of the selected algorithm to only those image objects (of the selected classes (step 2) on the selected image objects or image object level (step 1)) that fulfill the condition.
3.1 Click the broad No Condition b button (alternatively displaying the names of selected features) to open the Select Single Feature dialog box.
3.2 Select a feature to filter the image object domain for a selected feature. Choose a feature to open the Edit Threshold Condition dialog box, where you can define a threshold condition.
Edit Classification Filter on page 446
Select Single Feature on page 463
Features Overview on page 168
Figure 112: Click the broad buttons for modication of the items that are displayed on the button.
In the Feature group box, the feature selected to define the threshold is displayed on a broad button. To select a different feature, click this button to reopen the Select Single Feature dialog box.
3.3 In the Threshold settings group box, select a logical operator. Enter the number defining the threshold. With Definiens Developer, you can also select a variable, if one exists. For some features like constants, you can define the unit to be used. The feature range is displayed below the drop-down lists.
3.4 Click OK to close and return to the Edit Process dialog box. The defined feature threshold condition is now displayed on the broad button b.
Tip
Defining a threshold conditions is available within class descriptions too.
Use Thresholds for Classification with Class Descriptions on page 212
4. To specify the Image Object Domain, you additionally can define a Maximum number of image objects c.