Neural networks for uncertainty management in vision systems

Summary form only given, as follows. A novel methodology is examined for the fusion and propagation of uncertainties in a neural-network-like structure. Each node in the network represents a hypothesis. The inputs are the uncertainties associated with the knowledge sources that support the hypothesis, and the output is the aggregated uncertainty. The authors propose the use of fuzzy-set-based activation functions. The parameters of the activation function are chosen depending on the desired ´attitude´. In addition to the traditional union and intersection aggregation connectives, the authors propose the use of a generalized mean connective to increase flexibility. Some attractive properties of the connectives are discussed and a training procedure for such networks is proposed. Initial simulation results are also presented.<>