Narrow Results

Our objective in the interactive formulation of the "formal description schema - fds" model is the modeling of the prototypical, i.e. the subjective, perceptual ability of a human "expert", the ultimate human or robotic decision maker. In this paper, we present our fds-approach and methodology for solving the problem of modeling and exercising perceptual recognition [3–6]. We limit our discussion to one-dimensional variational profiles. We view the fds-model as a two-stage procedural model. Concerning the "early" (pre-attentive) recognition stage, we define the "structural identity of a k-norm class, k∈K" — SkID — as a tool for quick shadowing of sensory data and positioning instantiations of sufficient resemblance to interactively pre-defined spatio–temporal norm classes. Attentive recognition tools follow for assessing conformity of SkID-pointed occurrences.