ECAI-2002 Conference Paper

[full paper]

Deriving Textual Descriptions of Road Traffic Queues from Video Sequences

Ralf Gerber, Hans-Hellmut Nagel, Heiko Schreiber

Based on geometric results obtained by an algorithmic video (sequence) evaluation, a generic conceptual representation will be instantiated into a representation of the specific temporal developments within the recorded scene. Such an instantiated conceptual representation will in turn provide the input for a text generation subsystem. This contribution outlines a system implementation which relies on a fuzzy metric-temporal (Horn) logic to realize the system-internal generic representation. The instantiation step is realized as the search for a suitable interpretation of the corresponding set of logic formulae where the geometric results obtained by image sequence evaluation provide the set of individuals. The practicality of this approach is demonstrated by abstraction processes which aggregate suitably selected vehicles from recorded road traffic scenes into a vehicle queue.

Keywords: Vision, Knowledge Representation, Natural Language Processing, Qualitative Reasoning, Reasoning about Actions and Change, Knowledge-based Systems, Temporal Reasoning, Description Logics

Citation: Ralf Gerber, Hans-Hellmut Nagel, Heiko Schreiber: Deriving Textual Descriptions of Road Traffic Queues from Video Sequences. In F. van Harmelen (ed.): ECAI2002, Proceedings of the 15th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2002, pp.736-740.