Welcome to the Vid-Text-homepage of the 'Institut für Algorithmen und Kognitive Systeme' IAKS Nagel.

Vid-Text is an acronym for video-to-text-conversion. The links below try to cover most recent results in the automatic generation of conceptual and natural-language description of video sequences achieved within the CogViSys project.
The CogViSys project aims at the construction of an artificial cognitive vision system. Having this in mind, the field of video-to-text-conversion covers the evaluation of image cues and the transformation of those cues into conceptual descriptions and from there into natural language descriptions of video sequences. This process has to make use of knowledge about how image cues can be associated with primitive concepts. Other knowledge sources should be used which state the connections and relations between primitive concepts and more abstract or more complicated concepts.
Thus, appropriate representation formalisms for knowledge about the discourse in question have to be found and efficiently used.

MOTRIS MOTRIS is an acronym for MOdel-based TRacking in Image Sequences.
MOTRIS is, too, the name of our current system for detection and tracking of vehicles in innercity road traffic scenes.

SGTEditor The SGTEditor homepage. SGTEditor is a tool for graphical creation and inspection of situation graph trees, a representation formalism for behavioral knowledge.

F-Limette F-Limette is a logic inference engine for FMTHL - the fuzzy, metric temporal Horn logic developed by K.H. Schaefer in 1996.
With this inference engine and the underlying logic formalism, it is possible to provide a well-founded basis for conceptual video description.

Last edited:2005-02-17; For questions, suggestions or any other comment on this site please contact Vid-Text.