Introduction

The ImFusion Suite is a rapid prototyping software for working with medical images. It is built on top of the ImFusion SDK and therefore provides a rich and powerful toolbox of processing, analysis, and visualization algorithms. A plugin architecture allows for dynamically extending the set of functionality.

../_images/suite-intro.jpg

The main user interface is separated into several sections:

  1. The Data widget showing an overview of all loaded data.
  2. The views section showing 2D, MPR, and 3D visualization of the currently selected data.
  3. The Annotation Widget providing functionality to annotate images with elements such as labels, lines, etc.
  4. The Algorithm Dock providing space for the GUI elements of individual algorithms run on the loaded data.

Loading Data

All supported data formats can be loaded via the Import menu in the top menu bar. Usually this leads to a filesystem browser from which you can select the files to import. Certain data formats provide a dedicated GUI with additional options, such as the DICOM browser enabling you to scan entire directory structures recursively or connecting to a PACS instance on the network.

Alternatively, to the Import menu you can also use drag-and-drop to load existing data into the Suite. The software will use the file extension to determine the file format and forward it to the corresponding loader.

After importing all loaded data will be added to the Data widget (1). This GUI element serves two main purposes:

  • You can select one or multiple images that are to be shown in the views (indicated by the eye icon in the left-most column).
  • You can select one or multiple images and run algorithms on them (indicated by the highlighted background color).

By default, the visibility is synchronized with the current selection in the data list. You can disable this feature through the options menu in the top right corner of the Data widget and unchecking the Visibility Follows Selection option.

Views / Visualization

All images that are toggled as visible in the Data widget are automatically shown in the view section (2) of the Suite. Depending on its type/structure the images are distributed to one or multiple of the following view types:

  • 2D images and 2D image sets are shown in a dedicated 2D view.
  • 3D images and 3D image sets are shown in 3 orthogonal MPR views (slice rendering) as well as a 3D view (volume rendering).

More information about the views can be found in the dedicated page.

Annotations

The Annotations widget (3) enables you to place various annotation objects on top of images or other data, for example meshes. Annotations can be used for annotating purposes as well as for measuring.

More information about the annotations can be found in the dedicated page.

Running Algorithms on Data

The majority of the ImFusion Suite functionality is implemented through the concept of an Algorithm. These are modular building blocks operating on input data, performing processing and/or analysis tasks, and optionally providing new output data.

../_images/AlgorithmBrowser.png

In order to run an algorithm on one or multiple images you must first select them in the Data widget (1). For algorithms that do not need any input make sure that no items are selected. Then, you can use the Algorithms button in the menu bar to launch the Algorithm Browser that shows a list of all compatible and allows for filtering based on the name and documentation. Alternatively to launching the Algorithm Browser you can also directly right-click on the data in the Data widget to show a context menu listing the set of available algorithms that can be run on the selected input data. Selecting one entry will launch the corresponding algorithm.

Note

Some algorithms will show a small beaker icon next to their name indicating the experimental/research nature of the algorithm. This means that it may not be particularly polished and for instance only work for specific use cases or that it has not been tested with a wide variety of input data.

Some algorithms will run without further input from the user and directly provide the output/result. However, the majority of them will first spawn an new Controller UI that is shown in the algorithm dock on the left-hand side (4). Here you can configure algorithm-specific settings and evetually launch computation.

It is possible to run multiple algorithms and keep their controllers open at the same time.