Start Definiens Developer 32 Load and Manage Data 32 View Data in Projects 33 Develop Rule Sets 34 Process Data 35 Examine Results 35 Export Data 36 This chapter presents a overview of the workflow and sequences the basic operational
steps.
3.1 Start Definiens Developer
Start a Definiens client like Developer, Architect, Analyst, or Viewer together with an application to use specific functionalities for your specific tasks.
Industry- or user-specific applications enable users of clients like Definiens Architect to create ready-to-use solutions for their specific image analysis problem. Starting with other clients, an application completes the client functionalities by providing particular tools and controls needed for industry- or user-specific tasks.
Definiens offers a range of applications.
3.2 Load and Manage Data
Images can be analyzed in a single project or in a workspace which can contains several projects.
A single project is used to develop your image analysis solutions. A project refers to image data (called scenes). Image analysis extracts information from a scene and adds it to the project. Multiple projects can be put together and managed in a workspace.
A workspace is a file that bundles projects so that you can perform both manual and automated image analysis. In the Workspace window, you manage your image analysis tasks and administrate all relevant data. Thus, you control the workflow of automated image analysis.
Start Definiens Software on page 37
Load and Manage Data on page 43
Figure 19: Workspace window with Summary and Export Specification and drop-down view menu.
To run an automated image analysis you first have to import scenes to add image data to the workspace. In a workspace, each scene is administrated as a project.
3.3 View Data in Projects
You have several options for viewing and investigating your Definiens projects visually.
Some of the visualization tools are dependent on the extent of analysis that has been performed.
1. You can explore individual image layers right from the start, after loading scenes, and throughout the image analysis, including the investigation of results. You can define the color composition for the display of image layers and set equalizing options.
2. You can use different zoom and pan functions at any stage of image analysis.
3. After a basic segmentation that creates the first image objects, you can view their borders, called outlines.
4. Next, you classify image objects using various methods. You will commonly review the classification results extensively.
A specific visualization of features (image object attributes) enables you to preview the relevance of comprised information.
5. For refinement of your image analysis, you can navigate visually through the different (if there is more than one) levels of the image object hierarchy. This enables you to to maintain information about the different image object levels and to perform more classifications.
6. Additionally, you can use several views of a project at the same time.
View Data in Projects on page 94
View a Project on page 84 Use Workflow Views on page 40
Change Display of Image Layers on page 96
Zoom Within Projects on page 102
View Outlines of Image Objects on page 108
Change Display of Image Objects on page 109
Feature View on page 111
Navigate Within Projects on page 114
Use Different Viewing Modes on page 116
If you work with polygons you can display polygons and skeletons.
Figure 20: Project View: views of different image layers of a new project.
Figure 21: Project View: different views of a classified project.
Image data courtesy of Ministry of Environmental Affairs of Sachsen-Anhalt, Germany.
3.4 Develop Rule Sets
The process of detecting objects in an image consists of a sequence of segmentation and classification procedures. The sequence is defined in the rule set, which represents the program code. It consists of a list of processes, which represent commands, organized in a tree-like structure.
The typical workflow of the rule set development consists of three iterative activities:
• Use one the different ways of retrieving information about image objects.
• Edit one or more lines in the rule set.
• Run a branch of the rule set process tree for testing.
At the end you should have a rule set, that—when executed entirely on a image data—
detects the image objects of interest as defined by the image analysis task.
If you are already experienced in rule set development, you may continue with the extended workflow of rule set development focusing the strategic approach.
View Polygons on page 112 View Skeletons on page 114
About Image Analysis with Image Objects on page 124
Get Information About Image Objects on page 126
Use Processes to Build Rule Sets on page 130
Execute Processes for Testing on page 151
How to Approach Developing a Rule Set on page 326
Instructions About Developing Rule Sets
This User Guide provides three chapters about developing rule sets:
• Start Developing Rule Sets introduces the technology and the development workflow and explains the basic tools. This chapter enables you to create basic rule sets for simple image analysis tasks.
• Develop Efficient Rule Sets instructs you in using the available tools to build efficient and reusable rule sets and other ruleware. This chapter enables you to benefit from the tools of the development environment.
• About Development Strategies focuses on composing rule sets based on the Definiens Cognition Network Language (CNL). This chapter enables you to develop rule sets that suit your needs concerning both insight and efficiency and supports you in handling common problems.
3.5 Process Data
Processing applies ruleware, which may be a rule set or a solution, to the image data.
Thus, you receive image analysis results, and the results in turn enable you to evaluate the ruleware.
There are different methods for processing data:
• Analyze a single project by executing manually a rule set in the Process Tree window.
• Analyze a single project by executing manually a solution in the Analysis Builder window.
• Analyze single or multiple projects in a Workspace in an automated environment.
• When you analyze single projects with large image size, the image analysis maybe speed up by creating subsets or tiles of a larger project or image.
3.6 Examine Results
After processing image data you can examine the results, thus allowing you to evaluate the used rule set.
Various methods enable you to explore and evaluate the image analysis results. In addition, you may want to use manual editing tools to influence the results by hand.
View and investigate the results of an image analysis.
Results of an image analysis can be both numerical and visual. Numerical results are statistics and result values. Visual results are represented by the classification of image objects.
You can view statistics and compare numerical result values of projects. Depending on the analysis rule set, certain export functionalities enable the calculation of result values in relation to both single projects and large data sets of a workspaces.
Additionally, you can open a project to view and investigate it. A number of viewing options enable various visualizations of analyzed projects. Various display options enable in-depth exploration of both the image object hierarchy and image objects.
Start Developing Rule Sets on page 121
Develop Efficient Rule Sets on page 192
About Development Strategies on page 326
Process Data on page 363
Examine Results on page 382
3.6.1 Manual Editing
Manual editing of image objects allows you to manually modify the result of an image analysis. It can be applied to highlight or reclassify certain objects or to quickly improve the analysis result without re-editing the underlying rule set.
Image objects can be manually merged with neighboring objects, assigned to a selected class, or split into smaller objects.
Additionally, thematic objects can be manually created, merged, cut, deleted, or saved as a shapefile.
3.7 Export Data
You have several options for exporting data from the system in order to use it in other applications. Analysis result data can be exported as graphic information in vector or raster format. Additionally, statistical information can be created.
There are different basic methods to export analysis results:
• Data export triggered by a rule set that specifies where and how data is exported.
• Data export triggered by an action that specifies where and how data is exported.
• Data export initiated by Export menu commands. This is used to report data on a currently open project.
Manual Editing on page 394
Export Data on page 405
4 Start Definiens Developer
Contents in This Chapter
Start Definiens Software 37
Main Window 38 Workspace Window 40 Use Workflow Views 40 Handle Docking Windows 41 Customize the User Interface 42 This chapter provides an introduction for starting Definiens Developer. In addition, it
gives you an overview of the user interface.
4.1 Start Definiens Software
Start a Definiens client like Developer, Architect, Analyst, or Viewer with selectable portals offering predefined user interface arrangements.
A portal provides a selection of tools and user interface elements typically used for image analysis within an industry or science domain. However, most tools and user interface elements that are hidden by default are still available.
The following portals are available; the offered choice of portals depends on your Definiens eCognition Server edition:
Portal Description
Available only for Definiens eCognition EII Life Standard portal for the life sciences domain. • Life Server Cell Recommended for the cell based image analysis. Standard portal for Definiens
Cellenger application.
• Life Server Tissue Recommended for the tissue based image analysis. Standard portal for Definiens
Tissue Proliferation Assay application.
• Life Server EII Earth Standard portal for the remote sensing and geographic sciences domain. • Earth Server
1. Go the Windows Start menu and choose Start > All Programs >Definiens Developer > Definiens Developer.
Figure 22: Start a Definiens client from the Windows Start Menu.
A selection dialog box opens.
Figure 23: Selection dialog box offering different portals.
2. Select a portal.