Citation Context

...ime varying data. Aggarwal et al. [8] define ”activity/behavior” as a pattern derived from an image sequence. Cohn et al. 5Technion - Computer Science Department - Tehnical Report CIS-2009-06 - 2009 =-=[18]-=- define ”dynamic behavior” as a set of spatial states whose relationships are defined by temporal logic. Hongeng and Nevatia [19] describe ”simple events” as actions occurring in a linear time sequenc...

Citation Context

... would be interesting to compare the relevant aspects of their logic to this approach. 11. In addition to robotics, the theoretical framework of this article has been influential in cognitive vision (=-=Cohn, Magee, Galata, Hogg, & Hazarika, 2003-=-) and spatial reasoning (Hazarika & Cohn, 2002; Remolina & Kuipers, 2004). Acknowledgments M. Shanahan/Cognitive Science 29 (2005) 131 The work described in this article was supported by EPSRC grants ...

Citation Context

...sed on declarative models of activities. They define scenarios for Vandalism, Access forbidden and Holdup and use a hierarchy of facts ranging from abstract to concrete to recognize these situations. =-=[4]-=- investigates the use of qualitative 1 spatio-temporal representations and abduction in an architecture for Cognitive Vision. [11] considers an activity to be composed of action threads and recognizes...

Citation Context

...sed on declarative models of activities and defines scenarios for Vandalism, Access forbidden and Holdup and uses a hierarchy of facts ranging from abstract to concrete to recognize these situations. =-=[13]-=- investigates the use of qualitative spatiotemporal representations and abduction in an architecture for Cognitive Vision. [14] employs a context representation scheme for surveillance systems. [15] c...

Citation Context

...ometric properties of objects was further proposed within this system by (Schroeder and Neumann, 1996). Also based on the idea of sensor data assimilation as abduction, (Hazarika and Cohn, 2002) and (=-=Cohn et al., 2002-=-) propose a system for explaining video data. Following similar precepts, (Hazarika and Cohn, 2002) proposes a model for the assimilation of qualitative world descriptions from partial qualitative spa...

Citation Context

...tures as they are the result of a particular way of discretizing quantitative measurements into qualitatively interesting concepts, such that these concepts signify perceptually salient relationships =-=[3]-=-. The problem of abstracting qualitative relations from noisy video data, is facilitated by the use of a Hidden Markov Model based framework described in [15]. Five types of relations are illustrated ...

Citation Context

...orithmen und Kognitive Systeme, Fakultät für Informatik, Universität Karlsruhe (TH), Postfach 6980, D-76128 Karlsruhe, Germany. email: {ottlik|nagel}@iaks.uni-karlsruhe.de images and image sequence=-=s ([4], [1-=-5]). In this context, probabilistic methods like Hidden Markov Models, Bayesian Belief Networks, or Neural Networks (see, e.g., [2], [3]) constitute a ‘natural’ option for many researchers in the ...

Citation Context

...lexity of combined spatio-temporal knowledge representation formalisms [8] and investigates the use of qualitative spatiotemporal representations and abduction in an architecture for Cognitive Vision =-=[4]-=-. In the past cognitive systems, represented knowledge explicitly using chronicles [6] a temporal representation scheme for time, events and actions. Inspired by this approach, scenarios [13] have bee...

Citation Context

...cs and the interpretation rules are manually encoded into the surveillance systems, endowing them with the ability of "knowing" and "understanding", two of the three main features =-=of cognitive vision [5]. However,-=- considerably less research has been undertaken to provide surveillance systems with the third characteristic ability of cognitive vision systems: "learning". In previous research, the seman...