Output details
11 - Computer Science and Informatics
University of Leeds
From Video to RCC8: Exploiting a Distance Based Semantics to Stabilise the Interpretation of Mereotopological Relations
<22> Funded by a DARPA project (Mind's Eye, SRI subcontractor, $570k), this COSIT-11 Best Paper makes two major contributions: (1) a completely novel distance-based semantics for the ubiquitous qualitative spatial RCC-8 calculus (Cohn’s seminal KR-92 paper is GScholar >1700), rather than point-set intersection, allowing representation of, e.g., “nearly touching”; (2) a demonstration that more stable sequences of Qualitative Spatial Relations from noisy video data via a HMM based on the semantics thus improving activity recognition. The RedVine software (http://www.engineering.leeds.ac.uk/computing/research/vision/redvine/Redvine.htm) incorporates this technique which is now being exploited in two new EU projects on activity recognition (RACE, STRANDS: €1.5M).