Automatic semantic annotation Images and their parts are annotated with so called imagenotions. An imagenotion (formed from the words "image" and "notion") visually represents an entity, such as a person ("Nicole Kidman") or a movie ("Cold Mountain"). Such imagenotions can then be connected (e.g., the actress with the movie), multiple labels in different languages can be defined for them, and so on. Images and their parts are automatically annotated in ImageNotion to the maximum possible extent. We combine the results of object and face detection, face identification, and text mining to achieve the best results. Manual annotation is also possible, of course.

Visual exploration of archives Semantic metadata is used in innovative ways to provide a userfriendly, intuitive information system that supports the exploration of image archives. Because also image parts are annotated, a click on them can start a new semantic search or the current search can be extended. This allows a simple "click-through" navigation workflow for the end user and provides a novel, intuitive way to explore an image archive. Annotated image parts also make it easier to recognize the most important parts of the image and deliver useful background information.

Get ImageNotion ImageNotion is intended for professional image agencies, photographers and cultural heritage archives (museums, libraries). ImageNotion will be available soon in the following versions: Standalone The web application may be installed at the archive site, or a portal for multiple archives may be provided. Service Innovative parts of the system, such as the automatic generation of metadata, can be plugged into an existing image archive management system. Contact us

Start search application