Information
Deselect to remove the feature.
Select Features to Display Open the Select Displayed Features dialog box. Select Features on page 461
Find Out More
Display in Image Object Information
Another way to add or remove features is to right-click on a feature in windows, where you have the option to select features. For instance, in the Feature View window in the context menu select Display in Image Object Information and the feature appears in the Image Object Information window. Likewise, when editing or creating a class using the Insert Expression dialog box the newly inserted membership function will be added to the features selected within the Image Object Information window.
Edit Open the Edit [feature] dialog box with the same name as the feature. For example, if a feature from within the Standard Deviation feature tree group is selected, the Standard Deviation editing dialog box opens. It is used for editing parameter values of a feature.
In cases you select Edit or Create for a customized feature, you commonly activate the advanced Edit Customized Feature dialog box.
Create New Features on page 170
Edit Object Value Open the Enter Object Variable Value dialog box where you may change the selected image object value. The new value is displayed after its value in the feature tree as well as in the project view before the actual value.
Find Open the Find and Replace window. You may then substitute a created feature with another.
Find and Replace Rule Set Items on page 146 Delete Delete the feature completely.
Manage Variables Open the Manage Variables dialog box. Manage Variables on page 261
Manage Customized Features Open the Manage Customized Features dialog box. Create Customized Features on page 269
Review Results of a Classification on page 255
Save Save feature as .duf file.
Load Import saved customized features from .duf files.
Edit Unit Open the Select Unit dialog box. You change the feature unit if available. The unit of a feature is displayed after its value in the feature tree.
Not available for variables.
Save to File Save the dialog box contents as tabular .csv or .tcsv file.
Copy to Clipboard Save the active window to the clipboard. Copy Current View to Clipboard on page 413
7.5.3.1 Example: GetInfoOnImageObjects.dpr
For better understanding, we recommend opening the test project which is included within the installation. The test project is provided so that you might experiment with different methods and settings useful in analyzing information concerning image objects and features. We recommend saving your modified test project under a new name.
1. Open the test project by one of following ways:
• Open the project named GetInfoOnImageObjects.dpr. The path is:
C:\Program Files\Definiens Developer Version number\Examples
\TestProjects\GetInfoOnImageObjects.dpr.
• Alternatively, open the workspace named TestWorkspace.dpj. The path is:
C:\Program Files\Definiens Developer Version number\Examples
\TestWorkspace\TestWorkspace.dpj.
Next, open the Cell Assay folder within the Workspace tree view of the Workspace window.
Then view the project by double-clicking the Cells project within the Name column of the Workspace window.
2. Make sure the Image Object Information window is open.
3. You will notice that the image is present and a process tree is loaded and executed.
After clicking an image object in the project view, some features will appear in the Image Object Information window.
4. Go to the Image Object Information window and add the following features by right-clicking and choosing Select Features to Display on the context menu:
• Object features > Layer Values > Mean > Brightness
• Object features > Layer Values > Mean > Max. diff.
• Object features > Layer Values > Standard deviation > (all 3 image objects)
• Object features > Shape > Generic > Area
• Object features > Shape > Generic > Length/Width
• Object features > Shape > Generic > Roundness
• Object features > Texture > Texture after Haralick > GLCM Homogeneity >
All directions > (all 3 image objects)
5. Go to the project view and click on single image objects to compare their feature values listed in the Image Object Information window.
Open a Separate Project on page 43
Manage Projects in a Workspace on page 57
Image Object Information
6. In the Image Object Information window, double-click on Layer3:Nuclei. Now all image objects that are found within this layer are displayed in shades of gray. Each image object is displayed in a grayscale value according to the feature value that is selected for visualization. The low feature values are darker while the high values are brighter. Image objects displayed in red have not been defined for the evaluation of the chosen feature.
7. When the cursor is held above an image object for more than one second, the values of the image object will be displayed in the cursor tool tip.
8. In order to switch between the Feature View display method and the regular layer display in the project view, click the View Layer button or the Feature View button on the View Settings toolbar.
7.5.4 Compare Multiple Image Objects by Using the Image Object Table
Compare multiple classified image objects that are listed in table together with various feature values. Comparing multiple image objects will be done in different situations when you have already classified image objects that you want to investigate in detail.
You may want to compare image objects of selected classes in cases of:
• Evaluation of classifications.
• Selection of image objects according to sorted feature values.
• Debugging and quality assurance.
The Image Object Table window allows you to sort the image objects according to feature values.
1. To open the Image Object Table window do one of the following:
• Choose Image Objects > Image Object Table on the main menu bar
• Choose View > Windows > Image Object Table on the main menu bar 2. To configure the Custom tab of the Image Object Table window, right-click in the
text area, choose Configure Image Object Table in the context menu. The Configure Image Object Table dialog box opens.
Figure 140: Configure Image Object Table dialog box.
3. Select Classes to display all its image objects. The Select Classes for List dialog box opens. Select the classes individually by clicking them or using the move All buttons.
Get a Visual Overview of Feature Values on page 180
View Layer Feature View
Image Object Table
Select Classes on page 460
4. Additionally, you may include Unclassified Image Object by selecting the check box.
5. Select Features to display the feature values and the Select Features for Export dialog box opens. Select the features individually by double-clicking them.
6. Click OK to display all image objects and their feature values as configured in the Image Object Table. Each image object is assigned an ID number.
Figure 141: Image Object Table window.
7. You can click a column header for sorting rows according to column values.
8. Select an image object in the image or in the table to highlight the corresponding image object in the table or image respectively.
Figure 142: A selected image object is highlighted twice: in the Image Object Table window and in the project view.
9. Depending on the export definition of the used analysis, there may be other tabs listing dedicated data.
Context Menu Options
To display available context menu commands, right-click on the text area of the Image Object Table window.
Image Object Table Context menu commands
Usage
Enlarge Columns Expand every column to fit the width of the corresponding text.
Reduce Columns Minimize columns to default width.
Configure Image Object Table Open the Configure Image Object Table dialog box to select classes and features.
Select Features on page 461
Edit Object Annotation Open the Edit Annotation dialog box to add or modify annotations, which may be displayed in the Annotations column.
To display the Annotations column, select it as a feature under Object Features > Variables when you configure the Image Object Table.
Save to File Export the Image Object Table or a single entry as a .csv file, .txt file or .tcsv file.
7.5.5 Compare Feature Values by Using the Feature View
Compare image object attributes to evaluate which features are suitable for separating the relevant classes. A frequent task in image analysis is to reduce the number of features utilized for classification to the essential ones.
Feature View is a display method to visualize features over the entire project view. It helps the user to visualize the properties of image objects in a graphical way and therefore provides intuitive access to the peculiarity of a certain feature over all image objects in a project. Together with the Feature View window it allows you to perform the following tasks:
• Get visual overview of a feature by displaying an image object feature.
• Investigate threshold values of a feature by editing the range of the display of a feature.
The Feature View display method renders each image object in the project view according to its value for a selected feature. Each image object is displayed in a grayscale value according to the feature value that is selected for visualization. The low feature values are darker while the high values are brighter. Objects displayed in red have not been defined for the evaluation of the chosen feature. If you view, for example , the feature Rel. border to brighter neighbors all objects without a brighter neighbor will be displayed in red.
7.5.5.1 Get a Visual Overview of Feature Values
To get a visual overview of the properties of a specific feature, you can display an image object feature on the project view.
1. To display an image object feature, you have to activate the Feature View display method. Do one of the following:
• Go to the Feature View window and double-click a feature you want to visualize.
• Go to the Image Object Information window and double-click a feature you want to visualize.
• From the context menu of either window, use Select Features or Select Single Feature, and select the feature you want to visualize by right-clicking it and choosing Update Range on the context menu.
Each image object is displayed in a grayscale value according to the feature value
Get a Visual Overview of Feature Values on page 180
Investigate Threshold Values of a Feature on page 181
Select Single Feature on page 463
that is selected for visualization. The low feature values are darker while the high values are brighter. Image objects displayed in red have not been defined for the evaluation of the chosen feature.
2. Go to the project view and point (without clicking) to an image object to display its feature value in the cursor tool tip:
Figure 143: Using Feature View display method. (Image data courtesy of EMBL Heidelberg.)
In this example from the test project, the Feature View display method visualizes the ratio that the image objects have in the blue image layer. In this case you can see, that the selected feature (Object features > Layer values > Mean > Layer 3) is suitable for the separation of these image objects Nuclei. Darker image objects have a lower value than lighter ones.
3. Display another feature by double-clicking to compare it with the previous feature.
4. Click the View Layer button to stop the Feature View display method. Click the Feature View buttons to come back to the Feature View display method.