KAT:User Documentation

From Semantic Multimedia Wiki

Contents 1 Quick Start 1.1 Installing the KAT from a binary release 1.2 Installing from source 2 Default Plugins 3 Basic Workflow: Annotating Image Regions

Quick Start

Installing the KAT from a binary release

In order to run the KAT, you just have to download the latest release from Lauchpad and store it on your computer. You require Java6 to be installed.

Under Linux, change to the download folder and run

java -jar kat-<version>-bin.jar

The KAT should show up. If you had an older version installed on your machine and the KAT does not start correctly, try to run

java -jar kat-<version>-bin.jar --reset

once, and then the command above again. It might be the case, that some old preferences are still present in your Java backingstore.

If you use Windows, running the KAT is analogous, but a double click on the jar file should also work.

Installing from source

You need bazaar, maven and ant in order to checkout and build the KAT.

Just checkout a branch from the trunk series:

bzr branch lp:kat

Then change into the directory

cd kat

and run ant to build the KAT

ant

There will be an executable in the subfolder:

target/kat-<version>-jar-with-dependencies.jar

You can run this file as described above.

Default Plugins

There are four main plugins included in the KAT per default. On the left-hand side you will find the Ontology Browser, which is capable of loading OWL ontologies and displaying instances of ontology classes. In the middle area, there is the Image Annotation Tool, used to display and create regions within an image. The second plugin found in the middle is the Content Browser, which is used to show retrieval results and import content. In the top-right corner, you can see the Annotation Queue, which collects content that was analyzed by one of the analysis plugins. In the lower right corner the Annotation Browser is displayed, which gives a more structured view on the COMM annotations.

Basic Workflow: Annotating Image Regions

The basic workflow supported currently is the annotation of image regions. We will give a step-by-step description of this workflow:

1. Go to the content browser, click the folder icon and select an image to import.
2. If you have an analysis plugin, you could select the image, select the plugin in the drop-down on the bottom and submit the file. After the analysis has finished, the image will show up in the annotation queue.
   1. You can try the DemoPlugin, which will add a dummy region to the upper-left corner of the image.
   2. Or double click the image directly, which will open it in the image annotation tool.
3. Using the rectangle or polygon tools available from the Image Annotation Tool's toolbar, you can add regions to the image.
4. Open an OWL ontology in the ontology browser.
5. Select the shape selection tool (rectangle with cursor) and enter one of the regions. The region will highlight.
6. Now you can drag the region onto an ontology class, and the region will be annotated with an instance of the class.
7. You can see the COMM structure in the annotation browser in the lower-right corner.
8. Go to the content browser again, and click on the button with the small +. This will open a new tab.
9. Select the ontology class you used beforehand to annotate the region, and the image you annotated will show up.

Currently the KAT does not support deletion of annotation or regions, mainly due to limitations in the COMMAPI, but this functionality will follow soon.